01. Executive summary

We’re building a consulting business that enters on advisory — senior judgment the big firms can’t deliver AI-native — and then stays to run the function it recommends. AI and agents carry the production and the ongoing operation; a small senior team supplies the judgment, the accountability, and the relationship. Advisory is the tip of the spear; the embedded, software-run operating function is the body of the business. And we scale on software, not headcount — the moment we grow by adding people the way a traditional firm does, we’ve rebuilt the firms we’re displacing.

Who we serve

Regulated scale-ups and SMEs carrying a heavy compliance load they must meet, but that are too small to staff the function in-house and too small for a major firm to serve profitably — companies operating below the big players’ floor. The regulatory load is the forcing function: it makes a serious digital and compliance function non-negotiable exactly when the company can least afford to build one. This is deliberately guerrilla — we win where the incumbents can’t profitably follow. And it means the work is never about cutting staff: these clients can’t afford the headcount in the first place, so there is no one to displace — we add a function that wouldn’t otherwise exist.

What we sell

A working function on a subscription, not hours. Advisory gets us in the door and earns the right to operate; then the research, analysis, documentation, vendor evaluation, monitoring, and compliance tracking that consulting firms staff with juniors is carried by AI and agents at near-zero marginal cost — and so, increasingly, is the ongoing run. The senior human makes the build-versus-buy calls, negotiates with vendors, sits at the board table, and puts a name on every decision. Clients pay a monthly fee for an outcome they can see running. Part of what we leave embedded is an intelligence layer the client’s own people can run — their data, documents, and processes answerable in plain language, what dashboards always promised and never delivered. It is the wedge-to-managed-service motion the best cyber-security firms already run, applied to the digital operating function.

Why now

The pre-AI version of this model worked. One of the founders built and ran an embedded, retained function for a major APAC data centre operator in its scale-up years, and the engagement only ended because an acquisition created an audit conflict. The constraint back then was that every hour of the operating layer needed a person, which set a floor on viable client size and a ceiling on margin. AI removes that constraint — the operate layer can now run predominantly on software — which is precisely what lets us serve profitable functions for regulated clients sitting below the big firms’ cost floor.

The incumbents can’t follow. Big 4 and strategy-house economics rest on a pyramid of junior labour billed at multiples of cost. Most have also mis-sized the shift — pricing AI as a 5–10% efficiency gain when the reality is an order of magnitude, one person doing the work of ten — and even those who see it are responding by bolting AI onto the old advisory motion (a copilot to write the deck faster) rather than rebuilding advisory around agentic workflows. Doing what we do would mean dismantling the revenue engine their partnerships depend on. They are cutting graduate intakes, not restructuring. That leaves the position genuinely unoccupied.

Where we start

The segment is a pattern, not a sector, and it generalises across regulated industries. We enter where it is most acute and accessible right now: the cohort of Australian companies forming to build and operate the renewable grid — backed by infrastructure capital, facing demanding digital and regulatory obligations from day one, with no internal IT function. Sector selection is the venture’s core discipline, and Australian renewables went through the same gating evaluation we’d apply to any sector — ratified as the beachhead in June 2026 (section 10). The first play is the Renewable Energy Zone delivery cohort, where the founding team has a direct, currently embedded relationship. Renewables is the wedge; the regulated-and-can’t-afford pattern is the business (section 5).

The shape of the business

Fee and retainer based, with a small senior team running several embedded engagements at once because the production and operate layers no longer scale with headcount. The worked financial model (section 8) shows the practice profitable at five clients with three people, with margin expanding as clients are added without proportional hiring. Whether the venture ever takes equity in its clients is deliberately undecided; it isn’t an assumption the model depends on.

What we’re honest about

This plan names its open problems rather than smoothing them over. The load-bearing bet is the one in the model itself: how light the operate layer’s human footprint can genuinely go — whether it runs on software at scale or quietly reverts to bodies. Go-to-market is unproven. The quality ceiling — the point at which engagement complexity demands materially more senior time — is unknown until we run a real engagement, and the first one is designed to measure exactly that. The model scales only as far as senior judgment and trust scale. And the founding team’s transition out of current employment has constraints that shape sequencing. Each has its own section.

The founding team

Three founders envisaged: the owner plus two partners. Between them the venture needs senior client judgment and advisory experience, AI-native delivery capability (one prospective founder already runs a hybrid local + frontier model stack with routing in a live production system), and sector access (including a currently embedded position in the beachhead cohort). Roles, equity, and commitment are worked through together in section 14.

02. The shift: AI breaks consulting's production economics

Every consulting firm in the world runs on the same cost structure: every hour of output requires a person. An analyst costs $80,000 to $150,000 a year plus overhead whether they’re billable or not, so the whole industry is managed around utilisation, and growth always means hiring. The pyramid exists to arbitrage the gap between what a junior costs and what their work bills out at. David Maister formalised it decades ago: profit per partner is margin times productivity times leverage, where leverage is the ratio of juniors to seniors.

AI breaks the first premise. Output no longer requires a person.

What the production layer actually is

The work that consumes most junior and mid-level consultant time is structured production: decomposing a problem into testable branches, executing the analysis (the financial model, the market sizing, the competitor scan), and synthesising the result into a deliverable. In digital consulting the artefacts are even more mechanical: requirements documents, technology roadmaps, business cases, architecture options papers, vendor evaluations, governance packs, status reports. All of it is analytical production in a structured language, and all of it is exactly what AI agents now do well.

What AI cannot do is know whether the answer is right. A market sizing can pass every internal logic check and still rest on the wrong framing of what the client actually needs to decide. Code has automated tests; a strategy recommendation has no equivalent, so the validation mechanism for consulting output is senior judgment, and it can’t be automated away without removing the grounding for everything underneath it. That’s the durable human layer, and it’s what this venture staffs.

What the cost change looks like

The numbers are rough but the direction isn’t subtle. A junior analyst doing 40 hours of research costs a client roughly $6,000 to $12,000 through a traditional firm; the equivalent AI production run costs tens of dollars in compute. A full strategy engagement that a team of six delivers over three weeks at $250,000 to $500,000 has an AI production cost in the hundreds. Two orders of magnitude, sometimes three, and inference prices for a fixed level of capability have kept falling every year since 2022.

Senior time and the engineering effort to embed AI into a client’s real operations still cost real money, which is why the financial model in section 8 carries them honestly. But the production layer itself, the thing pyramids were built to staff, is now close to free.

What near-zero production cost changes

Utilisation stops mattering because there’s no standing army to keep busy; you pay for work consumed, not capacity waiting. Running ten investigative tracks in parallel becomes a decision about quality rather than budget, which matters because consulting problems are open-ended and the right framing is usually found by investigating widely, not by guessing well up front. Scaling decouples from hiring: more clients means more compute and a slightly fuller senior calendar, not an 18-month recruit-and-train lag. And pricing power moves to value. A cost base five to ten times lower first lets us serve clients profitably that the incumbents can’t reach at all — the big firms are simply too expensive for them, and the mid-tier, still heavily people-based, mostly aren’t built to — and then, where we do overlap, lets us price to the value delivered rather than an hourly rate while still out-margining a pyramid.

Why incumbents won’t simply do this too

The deepest reason isn’t that they can’t act — it’s that most of them have mis-sized what’s happening. The prevailing view inside the big firms is that AI is a five-to-ten-percent productivity gain for their consultants: a faster first draft, a quicker deck, the same work done a little sooner. We think that’s wrong by an order of magnitude. This isn’t ten percent off the cost of a deck — it’s one person doing the work of ten, and the production layer ceasing to be a staffing problem at all. A firm that treats a step-change as an efficiency tweak is optimising the wrong thing entirely.

And the firms that do grasp the magnitude are trapped anyway. Responding properly means dismantling the pyramid that pays their current partners, so the rational short-term move is to layer AI onto the existing structure and protect the revenue engine. That is what’s happening: the major firms are cutting graduate intakes (reported reductions of roughly 30% at PwC and KPMG, 18% at Deloitte, 11% at EY) while leaving their economics intact. Cutting the junior layer without changing the model shrinks the pyramid; it doesn’t replace it.

The window this opens has a clock on it. Industry analysis puts an 18-month horizon on providers demonstrating genuine AI-native delivery before they’re excluded from new mandates, with full market repricing of consulting fees playing out over several years. New entrants who are AI-native from the first engagement carry none of the structure that stops incumbents from acting, and the first credible players in each niche will set the reference point clients use to judge everyone else.

03. The model: advisory-native AI, built to operate

Traditional consulting sells capacity: a team, a timesheet, and a report at the end. We sell judgment — and then we stay to make it real. Advisory is how we enter: a senior partner at the table who has actually done the work, not a graduate with a framework. But unlike a consulting firm, we don’t stop at the recommendation; and unlike a software vendor, we don’t drop a tool and leave. We advise, we build, and then we run the function — and the part that runs is carried predominantly by software and agents, not by an ever-growing team of people. Advisory is the tip of the spear. The embedded, AI-run operating function is the body of the business.

Advisory is the way in — and where our edge is real

The thing we have that a frontier AI lab and a born-yesterday delivery startup do not is earned advisory judgment: founders who have sat in the room, made the call, and carried the consequence across large, regulated programs. That is the scarce input, and it is the hardest thing to fake.

What we don’t do is what the big firms are doing with it. Watch closely and the incumbents — ours included — are bolting AI onto the old advisory motion: a copilot to write the deck a little faster, a notebook to summarise the interviews. The process underneath is unchanged. The opportunity is the opposite — rebuild advisory itself around agentic workflows, so the work is done differently, not just faster. That is what we mean by advisory-native AI, and it is wide open: the firms with the judgment are too invested in the billable hour to cannibalise it, and the firms taking AI seriously don’t have the judgment.

Our buyer is a company under heavy regulatory load that is too small to staff the function properly and too small for a major firm to serve profitably — a regulated scale-up or SME operating below the big players’ floor. The regulatory load is the forcing function: it makes a serious digital and compliance function non-negotiable, exactly when the company can least afford to build one. Renewable-energy delivery entities are where that need is most acute right now (section 5), but the pattern is the segment, not the sector. This is deliberately guerrilla — we win where the incumbents can’t profitably follow.

From day one such a client gets an accountable technology executive at the board table, an architecture designed for where the company will be in three years, vendor selection run by someone who has done it at scale, and systems implemented, integrated, and operated. They don’t get a deck recommending these things — they get the things.

Then we stay — and we scale on software, not people

A recommendation a client puts out to tender is a commodity. So we don’t hand the work over at the boundary; we embed and operate the function we designed. This is where the business actually lives — the recurring, sticky, long-term layer, priced as a subscription to a working function rather than as hours.

And it runs on one hard principle: the operate layer is carried predominantly by software and agents, with the lightest human footprint we can hold. If we scale by adding people the way a traditional firm does, we have simply rebuilt the Big 4 with better marketing — the same economics, the same ceiling. We scale on software. Headcount is the thing we are deliberately engineering out of the growth curve.

We hold the degree as a goal, not a solved problem: how much of the operating layer genuinely runs on agents, versus how much still needs a hand on it, is the open question the first engagements exist to answer (section 8). But the direction is not negotiable — every part of the function we can move from people to software, we move.

We make people more capable — we don’t replace them

To be plain about what this venture is not: we are not here to walk into an organisation and cut a third of its staff. That’s deliberate, on three counts. It isn’t the company we want to build. We also don’t believe wholesale AI-for-headcount replacement is real at this scale — that story is oversold. And to the extent any organisation does want pure workforce reduction, that’s an enterprise game, served by players built for it; it isn’t ours. Our clients are the mirror image: they can’t hire the people, can’t afford the expertise, or can only do this badly or not at all. There is no one to displace — we add a function that otherwise wouldn’t exist. The segment doesn’t merely make that stance affordable to hold; it makes it true.

The same instinct runs through delivery. The operate layer is not a black box we guard — part of the job is lifting the client’s own people onto the same agentic tools and workflows, driving adoption inside their business rather than hoarding it inside ours. That compounds three ways: the client extracts more value and leans on the function more, so it grows with them; the relationship gets stickier because the workflows are ours and now run through their operation; and it keeps us on the right side of the line — we make people more capable, never quietly replace them.

The cyber-security parallel

This is not a novel shape; it is the model the best cyber-security businesses already run. They enter on advisory — a security assessment, a virtual CISO — to earn trust and learn the environment, then land the recurring managed service, where a platform and a small expert team detect and respond around the clock at software margins. The assessment is the wedge; the managed service is the business. A string of these companies have been built and sold on exactly this motion — several acquired by the big consulting firms themselves. We are running it for the digital operating function rather than the security function, with agentic systems as our platform.

The three roles

The team that does this is an inversion of the consulting pyramid — three senior-weighted roles, none of them a billing-fodder layer:

• Client Leader — the senior executive (Director/MD-calibre) who owns the relationship, sits with the board, makes the build-versus-buy calls, and puts their name on every decision. The advisory judgment, and the accountability. The scarce resource the business is measured in.
• Engagement Architect — the senior delivery technologist who designs the systems, stands up implementations and vendors, and turns decisions into working architecture across several clients at once.
• AI Facilitator — the operator of the software layer: building, directing, and quality-controlling the agentic systems that carry the production and the ongoing run — research, analysis, documentation, monitoring, compliance tracking.

A three-person practice with these roles serves five clients in the financial model (section 8); nine people serve twenty. The pyramid is gone because the work the pyramid used to do is carried by software.

The leverage equation, rewritten

Maister’s classic formula — profit per partner equals margin × productivity × leverage — defined leverage as the ratio of juniors to seniors, which is why every traditional firm’s growth plan is a hiring plan. In this model leverage is the ratio of AI-carried work to senior time. A Client Leader in a traditional firm carries two to four clients because production drags on their team; here the working assumption is six to eight, because production and run don’t touch them. This is the model’s most load-bearing assumption: section 8 shows what the economics look like if it proves to be four rather than eight, and validating it is an explicit objective of the first engagements. It is the same question as “how light can the operate layer’s human footprint really go” — asked from the financial side.

Why a better model doesn’t erase this

The obvious challenge to any AI-native firm is that the next, smarter model simply does the work itself. It doesn’t — because the defensible ground was never the model. A more capable model still doesn’t hold the client’s operating licence, sign off on the liability when the regulator calls, or own the private files, systems, and context the work runs on. That ground is private and accountable, and it is exactly where we embed: inside the client’s regulated business, with a senior human whose name is on the decision. Better models make our software layer cheaper and sharper — they don’t reach the corner we occupy. The scarce thing was never the capability; it is knowing what to point it at, and being answerable for the result.

What AI does, and what it can’t

AI carries the production and, increasingly, the run: market and vendor intelligence, architecture options, specification and documentation, compliance tracking and regulatory-change monitoring, system monitoring and alerting, board-pack preparation. This is the work consulting juniors do today, delivered at a cost measured in dollars rather than salaries (section 2 sets out that economics).

AI cannot know whether the answer is right for this client, this board, this regulator. It cannot negotiate a vendor contract, read a founder’s risk appetite, or be accountable when the market operator calls. Every engagement runs on a simple division: AI produces and runs, seniors decide, and the client always has a name on the decision.

04. How organisations deploy AI — and where we sit

Most companies think “using AI” means having a ChatGPT or Claude subscription. That is one point on a spectrum. For a regulated operator the real decision is not whether to use AI — it is where each workload should run, and who is accountable when it does. Routine drafting can sit with a public model; a board pack built on confidential financials, or an operational process touching a regulated system, cannot. The same company will, and should, sit at several points on this spectrum at once.

The spectrum, plainly

Read left to right, each step trades convenience for control. The cost of running a model falls and your sovereignty over the data rises — but so does the operational burden you take on.

Two things make this our business, not just a chart

Model routing is the layer, not a rung. At every tier you still choose which model runs which task — a frontier model for judgment and synthesis, a cheaper or specialised one for routine, structured work — and the routing layer makes that call automatically on cost, privacy and capability. Owning that layer for a client is the defensible position: it holds the accumulated judgement of which task goes where, and it lets a client move a workload up or down the spectrum without rebuilding anything. It is the same embedded, hard-to-rip-out ground the rest of this plan describes — applied to the plumbing of how AI actually gets used.

You don’t build the bunker on day one. The right posture is to start convenient and move workloads toward control only as sensitivity demands: begin at Tier 0 or 1, move regulated data into Level 2, and reserve Level 4 for the functions that genuinely must survive disruption. The hybrid pattern — a local or in-tenancy model processes the raw, sensitive material and only a structured summary ever reaches a public frontier model — is not a cost trick. For a client under regulatory load it is a compliance architecture: the evidence that sensitive data never left the perimeter.

Where we sit

This reframes what we sell. Not “we’ll help you use AI” — but “we’ll decide where each of your workloads runs, build it there, and own the routing and sovereignty layer that keeps you in control.” For the regulated scale-ups this venture targets, that is precisely the capability they cannot staff and the frontier labs will not provide: an accountable operator who places the work along this spectrum on their behalf, keeps their data on the right side of the line, and is answerable when a regulator asks. It is the moat in section 7, expressed as architecture — and it directly addresses the model-portability and sovereignty exposure the venture has to manage in its own delivery.

05. Why now: the timing window

Timing arguments are cheap, so this one is specific. Three independent clocks have to read the same time for this venture to work: the technology has to be able to carry consulting production, the beachhead market has to be forming, and the incumbents have to be unable to respond. All three are aligned now, and each one moves.

Clock one: AI crossed the production threshold

Until recently, AI could draft fragments of consulting work; now it carries whole production workflows. The current generation of frontier models one-shots tasks that were previously team-scale — the most cited public example is a 50-million-line code migration completed in a day that had been scoped at two team-months — and the agentic pattern of long-running, self-checking work loops has moved from demos into production at the companies that build these systems. The production layer of consulting (research, analysis, synthesis, documentation, monitoring) is precisely the work this generation does well, at a cost of dollars per deliverable rather than salaries.

The capability proof matters because the alternative explanations have run out. When an ASX-listed property group puts an AI agent over its SAP estate and 700 staff use it in production within 18 weeks of business-case approval, the question “is the technology ready?” is settled for the class of work we’re proposing to do.

And the threshold isn’t only that AI can carry production — it’s that agentic workflows now let the advisory process itself be rebuilt, not just sped up. This is the move the incumbents are not making: they bolt a copilot onto the existing motion to write the deck a little faster, leaving the process underneath unchanged. AI-native advisory — redesigning what the work is — is open precisely because the firms with the judgment are treating AI as an accelerant rather than a redesign.

The proof is loudest in law — the regulated profession furthest down this path. Harvey reached an ~$11B valuation delivering AI-native legal work at premium prices, and Legora hit $100M in annual recurring revenue inside 18 months — evidence that AI-native professional delivery commands value, not discount pricing, even in a liability-heavy field. The same curve is coming for the rest of professional services, including the work we do.

Clock two: the cohort is forming now

The companies our beachhead serves are being created in this window, not at some future point. Renewables passed 51% of Australia’s National Electricity Market in late 2025; the market operator’s 2026 system plan locks renewables, storage and transmission in as the build-out path to 2050; and the Renewable Energy Zone programme is bringing newly formed delivery entities into existence with hard digital obligations from day one. Industry practitioners describe 2026 as the year operators stop treating digital as an innovation project and start running it as their primary lever.

Formation timing is the wedge’s advantage. Greenfield is where AI-native delivery is most effective (the documented productivity gap between greenfield startup contexts and brownfield enterprise environments is roughly 100× versus 10%), and the embedded position is only available before a company builds an internal function. Once this cohort matures, the entry changes from “be the function” to “displace one”, and the economics and sales motion both degrade. The broader segment — regulated organisations that can’t afford to staff their compliance load — is not time-limited in the same way; but the renewables wedge, the cohort forming right now, is.

Clock three: incumbents are paralysed, not asleep

Most incumbents have mis-sized what’s happening — pricing AI as a five-to-ten-percent efficiency gain rather than the order-of-magnitude shift it is (section 2). And the minority who do grasp the magnitude can’t respond without dismantling the pyramid that pays their partners, so their observable behaviour is margin protection: graduate intakes cut roughly 30% at PwC and KPMG, 18% at Deloitte, 11% at EY, while delivery economics stay intact. AI is being layered onto their model, not allowed to replace it.

The sharpest magnitude signal comes from the top of the adjacent profession. The world’s largest law firm, Kirkland & Ellis ($10.6B revenue), is reportedly spending around $500M building its own internal AI platform — its chairman’s framing is that commodity tools have “raised the floor for everyone, but we don’t get hired for the floor.” That is not a firm that thinks this is a 5–10% tweak. But a half-billion-dollar self-build is an option only the giants have: the regulated scale-ups we serve can’t build it, can’t buy it at that scale, and can’t hire it — which is precisely the opening.

This is the textbook structural trap, and it does not last forever: the analyst who named the services-as-software category puts an 18-month horizon on providers demonstrating real AI-native delivery before being excluded from new mandates, and credible AI-native entrants are already forming overseas — Unity Advisory raised US$300M to do senior-only, AI-native CFO advisory; Distyl reached a US$1.8B valuation on AI-native enterprise delivery.

What closes the window

Three things, on different timescales. The cohort window closes as beachhead companies build internal teams — a 2026–2028 phenomenon. The differentiation window closes as “AI-native” stops being a distinction and becomes table stakes — the 18-month repricing clock. And the cost arbitrage narrows as subsidised inference normalises: enterprises now actively manage AI spend (98% of them, up from 31% two years ago), and the cost advantage will accrue to firms that engineered for token efficiency from the start rather than bolting AI onto old delivery.

None of these clocks says “wait and see”. Every quarter of delay surrenders formation-stage clients to whoever moves first, and the referral dynamics in a small market mean the first credible player sets the reference point. The honest tension — moving fast against the founding team’s employment constraints — is treated explicitly in the roadmap rather than wished away.

06. Beachhead: Australian renewable energy

Our segment is a pattern, not a sector: regulated scale-ups and SMEs carrying a heavy compliance load they must meet, but are too small to staff for — and too small for a major firm to serve profitably. We enter where that pattern is most acute and most accessible right now: the new cohort of Australian companies building and operating the renewable grid. Renewables is the wedge; the regulated-and-can’t-afford pattern is the business.

Australia is replacing a grid built around large coal and gas generators with a distributed system of renewable generation, storage, and grid services. Renewables supplied 51% of the National Electricity Market in the December quarter of 2025, per AEMO’s 2026 Integrated System Plan. This isn’t a forecast we’re betting on; it’s the operating reality the market hit last year.

That transition is creating an entirely new cohort of companies: utility-scale solar and wind generators, battery storage (BESS) operators, virtual power plant aggregators, grid-service providers, energy retailers built on the new infrastructure, and fund-backed asset operators. They share three traits that matter to us. They have capital, typically from infrastructure funds, private equity, or climate-focused investors. They face hard operational and regulatory obligations from their first day in the market. And they have no internal IT function, and no time to recruit one before they need to be operational. Funded, regulated from day one, and unstaffed — that is the pattern in its sharpest form.

What these companies have to stand up

Participating in the Australian energy market is digitally demanding in a way most early-stage businesses never face:

• Market participation. AEMO registration and interface systems, real-time bidding into the NEM, forecasting of generation, demand, and price, settlement and billing. AEMO is a mandatory interface, not an optional one; registration, dispatch, and telemetry obligations are conditions of being in the market at all.
• Operational technology. SCADA for monitoring and controlling physical assets, OT-to-IT integration, telemetry pipelines from the assets themselves. Cloud SCADA is already mainstream here; one Siemens deployment covers eight assets across four states at around 300,000 data tags.
• Compliance and security. Clean Energy Regulator obligations for renewable certificates, AEMC regulatory requirements, and cybersecurity rules for critical infrastructure, which the 2026 NEM reforms strengthen further alongside new consumer-energy-resource integration rules and expanded AEMO powers.
• Corporate systems. ERP configured for project-based energy asset accounting, PPA contract management, investor reporting.

Setting this up isn’t an IT task you hand to a contractor. It needs someone who can make executive-level build-versus-buy and vendor decisions, design the architecture so it survives the company’s growth, and then actually run it. That combination is what we sell.

Why the window is now

Companies formed between 2024 and 2027 to build and operate the new grid are making their foundational technology decisions in this period, and those decisions are load-bearing: expensive to change later, and formative for everything built on top. The firm embedded at that moment becomes the architecture’s author rather than a later vendor competing against an installed base. Industry practitioners called 2026 the year operators stop treating digital as an innovation project and start running it as their main decarbonisation lever, which matches what the regulatory calendar is forcing anyway.

Once this cohort matures and builds internal teams, the entry point changes from “be the function” to “displace an existing function”. That’s a much worse position, and it’s why we treat the window as time-limited.

Why Australia specifically

The infrastructure-investor community here is small, interconnected, and referral-dense; one well-run embedded engagement is visible to the whole cohort’s backers. Regulatory complexity (AEMO, NEM, CER, AEMC) rewards specialised knowledge and punishes generalist IT consultancies, which builds a moat that compounds with each engagement. Government policy at federal and state level is actively funding the transition, so the demand isn’t speculative. And the founders’ data-centre scale-up years were spent in exactly this Australian infrastructure world; the network reaches the people who back these companies.

Why the position is empty

This is the guerrilla logic of the whole venture, sharpest here. Big 4 economics need engagement sizes these clients can’t sustain, and their pyramids can’t profitably serve a three-person company that needs a working function rather than a report. Strategy houses don’t operate in this segment at all. Traditional IT consultancies deliver a project and leave; managed-service providers can run existing infrastructure but can’t architect it or advise a board. We win below the incumbents’ cost floor — the ground they can’t profitably reach — and the position requires executive technology judgment, operational execution, and an AI-carried production layer in one offer, which nobody currently occupies.

The evaluation result

Australian renewables went through the venture’s full gating evaluation in June 2026 and was ratified as the beachhead — clearing all five gating criteria with no disqualifiers (section 10 presents the worked, evidence-based assessment). The structural criteria scored highest: day-one digital complexity and the senior-judgment requirement are conditions of NEM participation, not assumptions. The scores held back from 5 are honest evidence gaps — the cohort and competitive field are structurally inferred but not yet census-mapped — and closing them is the go-to-market research now underway.

The evaluation also sequenced the entry. First play: Renewable Energy Zone delivery entities — the newly formed consortia and network operators standing up REZ transmission and enabling infrastructure, where the founding team holds a direct, currently embedded relationship. Battery storage operators are the second motion, utility-scale generation platforms the third; VPP aggregators were parked as too digital-native to need an external function.

Why this is a wedge, not a one-sector bet

Renewables is the opportunity that exists now, but the model is built to travel. The same pattern — a funded, heavily regulated organisation that must run a serious function before it can afford to staff one — recurs across regulated industries: healthcare, financial services, aged care, and critical infrastructure more broadly. The renewables beachhead is where we prove the model, the playbook, and the AI-carried operate layer; the segment behind it is far larger than one sector, which is what turns a niche entry into a venture.

The founding team has seen this shape before. A major APAC data-centre operator, in its scale-up years, faced exactly it: complex, heavily regulated, scaling faster than it could conventionally staff, and served instead by an embedded, senior-led function rather than a traditional consulting engagement. That worked. What’s new is the AI-carried operate layer that lets the same pattern be served at software economics — and pointed at the regulated mid-market, not just the well-capitalised giant.

07. The engagement: what we actually do

The engagement runs in four moves: land on advisory, embed to build, operate on software, and expand as the client grows. Advisory is how we earn the right to the rest — a senior partner who has done the work, not a recommendations deck — but unlike a consulting firm we don’t stop at the advice, and unlike a software vendor we don’t drop a tool and leave. It is the wedge-to-managed-service motion the best cyber-security firms run: the assessment earns trust and maps the ground; the recurring operated function is the business.

Land: the advisory wedge

We enter on judgment. The sharpest entry point is the moment a regulated scale-up’s compliance load steps up faster than it can hire — a funding round, a licence threshold, a new market-participant obligation — and a senior advisor who has navigated exactly that before is both the cheapest thing to say yes to and the hardest to fake. The advisory engagement does real work (the architecture call, the build-versus-buy decision, the regulatory path) and, in doing it, earns trust and a deep read of the business. That read is what makes the next move possible — and it’s something a frontier model and a delivery-only startup can’t bring.

Embed: build the function

A Client Leader acts as the company’s CIO/CTO: board attendance, technology strategy, build-versus-buy decisions made with the CEO and CFO, accountability to the board. The Engagement Architect designs the architecture from scratch and delivers it — for our beachhead: AEMO market-interface systems, SCADA/OT integration, compliance tooling, ERP configured for project-based asset accounting, contract and investor-reporting systems, security posture and IT governance. Formation-stage technology decisions are load-bearing — expensive to change later and formative for everything built on top — which is exactly why they need executive judgment, not a junior team and a methodology.

One pattern we build in from the start: the agent-over-systems interface. The strongest current Australian enterprise example put an AI agent over SAP so 700 occasional users could query procurement data conversationally inside Teams — live in production 18 weeks from approval. Clients’ systems of record arrive with a natural-language layer over them, so a lean team’s data actually gets used.

Operate: run it on software, not people

This is where the business lives — the recurring, sticky, long-term layer, priced as a subscription to a working function rather than as hours. And it runs on one principle: the operate layer is carried predominantly by software and agents, with the lightest human footprint we can hold. Vendor scans and market intelligence in days; specifications and governance packs produced as systems are designed; compliance obligations and regulatory change tracked continuously; monitoring and alerting from day one of live operation; board packs and investor reporting drafted for senior review. If we ran this layer with people the way a traditional firm does, we’d have rebuilt the Big 4 — same economics, same ceiling.

How light that footprint can genuinely go is the model’s load-bearing question, and we treat it as something to prove, not assert. The first engagement is scoped to turn the principle into a number: take one operating function, run it on agents, and measure it — where senior time actually goes, where AI output needed correction, and at what complexity the quality ceiling appears (sections 8 and 10). The direction is non-negotiable; the degree is what the first engagements exist to establish.

Empowering the client: from how we work to how they work

Capability transfer runs through every engagement, in two moves. First, during delivery, we run the engagement itself on AI in the open — showing the client’s people how we use it, so the way we work becomes the way they work. Then we flip the script: we embed AI into the organisation’s own infrastructure so they can run the business more effectively without us in the loop for everything.

The flagship of that embedded layer is the organisation’s intelligence layer — the same architecture the founders already run day to day, turned around and put inside the client. It puts an organisation’s scattered data, documents and processes in one place and makes them answerable in plain language: a CFO asking “how do our monthly financials look?”; a project lead asking “what’s actually happening on this build?”; a CEO treating strategy as a living document — “here’s our strategy, here’s our procurement process; tell me where we are and what I need to do next.” It is what dashboards always promised and organisations almost never delivered — in effect the “collective intelligence of the institution” that the world’s largest law firm is spending ~$500M to build for itself (section 4), delivered to organisations that could never fund that build. The value isn’t a chatbot bolted on the side — it’s putting the right information in front of the right person, building the processes that keep it true, and doing it to a standard a regulated organisation can actually rely on. That reliability bar is exactly what a bought tool or a DIY attempt doesn’t reach, and it’s why this is a delivered, supported capability rather than a download.

This is also where the relationship compounds and the line holds: the workflows are ours and now run through their operation (stickiness), the client’s own people get more capable (adoption), and at no point is the pitch “replace your staff” — there is no one to replace, only a function being built.

Expand: grow with the client

As the client builds internal capability, the engagement deliberately narrows rather than clinging to scope. Day-to-day operations transfer to the internal team we helped hire, onboard, and train — building our own partial replacement is part of the service, not a threat to it. What stays: the technology roadmap, the major calls (new asset classes, acquisitions, market expansion), vendor and regulator relationships at the senior level, and the AI-powered intelligence layer — market monitoring, regulatory-change tracking, competitive signals — that is cheaper for us to run across many clients than for any one client to replicate. The relationship rides the client’s growth: we land small, below the incumbents’ floor, and grow with them, which is what keeps a guerrilla entry from being a ceiling.

This is the most durable relationship shape in professional services: formed when the client had nothing, deepened by the fact that we designed everything they run on. The switching cost at year three — after we’ve built their systems, embedded their governance, and trained their staff — is not contractual; it’s structural. And it’s the untrainable corner a better model can’t reach: a smarter model still doesn’t hold the licence, sign off on the liability, or own the client’s files and systems.

What we deliberately don’t do

We don’t do bodies-on-seats augmentation, one-off projects that end at a handover, or work where the production layer is physical rather than digital. And we cap embedded engagements per Client Leader rather than stretching the senior layer thin — the model’s economics rest on the quality of senior judgment, and the fastest way to destroy the franchise would be to dilute it. The capacity assumption (six to eight clients per Client Leader) is tested explicitly in the first engagements, with the instrumentation built in to measure it.

08. Competition and why incumbents can't follow

The competitive question for this venture isn’t “who else does this?” — today, nobody does. The question is why the position stays empty long enough for us to take it, and what happens when others notice. Both have structural answers.

The map

The position requires four things at once — executive technology judgment, operational execution, an AI-carried production layer, and a commercial structure shaped around small clients that grow. Each competitor class has some; none has all four, and for the incumbents the missing pieces are missing for structural reasons, not because nobody thought of it.

Why the incumbents stay paralysed

This is Christensen’s incumbent dilemma in its cleanest form. The Big 4 and strategy houses earn their margins from leverage — junior labour billed at multiples of cost. Serving our market would require senior-only teams, sub-$500K engagements, multi-year embeds, and AI replacing the billable layer: every element is margin-dilutive to their current model, and the partners who would have to approve it are the ones whose income it would cannibalise. Their own client research shows the strain (one major strategy firm’s survey found 60% of clients getting no material value from AI engagements), and their observable response — protect the engine, trim the intake — is rational for them and convenient for us. The “fat smoker” problem: they know what to do; their structure stops them doing it.

There is a subtler version of the same trap, and it’s the one that matters most for our differentiation. Even where the incumbents are adopting AI, they are bolting it onto the old advisory motion — a copilot to draft the deck faster, a notebook to summarise the interviews — rather than rebuilding advisory around agentic workflows. The process underneath is unchanged; the AI is an accelerant on a model designed for billable hours. They can’t do otherwise: genuinely AI-native advisory collapses the very hours they sell. So the gap isn’t only that they won’t serve small regulated clients — it’s that they aren’t building AI-native advisory at all, in any segment. That is the ground we’re built on, and it is wide open.

There’s a third reason, less visible from outside the industry but decisive. The big firms have thousands of people to keep utilised, which pushes them relentlessly toward the large prize — the hundred-million-dollar ERP and Salesforce implementations. In that machine, advisory is a feeder, not the product: valued because it leads to the big implementation-and-operate deal that someone downstream actually wins and is celebrated for. Advisory-led, embedded delivery for a sub-floor regulated client is the opposite of what those economics are built to chase — inside a big firm, that work is a hiding to nothing. The mid-tier is no closer: still heavily people-based and, as best we can see, barely engaging with AI-native delivery at all. So the segment isn’t merely unprofitable for the incumbents to serve — advisory-led embedded delivery is structurally unattractive to them even where they could reach it.

The validating analogues

The strongest evidence the model works is that adjacent versions of it are working at scale, in different niches:

• Unity Advisory — US$300M from Warburg Pincus (March 2025) for senior-only, AI-native, embedded CFO-office advisory to PE-backed mid-market companies, founded by PwC UK’s former managing partner. The same structural bet — senior judgment over AI production, embedded and continuous — one function over and one market segment up from ours. It validates that institutional capital backs this model; it doesn’t touch formation-stage companies or the technology function.
• Harvey — US$11B valuation, ~US$1,200/seat/month, the proof that AI-native delivery of professional work commands premium pricing at enterprise scale in a regulated profession (law). A software company, not a competitor — and evidence for our pricing argument in section 9.
• Distyl — US$1.8B valuation for AI-native enterprise delivery built by ex-Palantir engineers: the forward-deployed pattern, commercialised. Fortune-500 horizontal, not formation-stage vertical.
• Future Secure AI — the closest structural analogue: Australian-founded, now Austin-headquartered, ~300 staff and tripling, selling embedded, operated AI “co-workers” to Global 2000 enterprises on five-to-ten-year lifecycles under an explicitly “not SaaS, not consulting” partnership model. It proves an AU-origin embedded-AI-delivery firm can reach enterprise scale — and it defines our segmentation by contrast: they serve the big end of town and left Australia to do it; we serve formation-stage entities in regulated niches, deliberately AU-sovereign, at a client size their cost base can never profitably reach. One deliberate divergence: they front their agents with AI-generated human personas; we embed agents as named functions with the human partners carrying all client trust.
• Robots & Pencils — a North American digital consultancy reinvented “AI-first,” marketing itself as the “nimble, high-velocity alternative to traditional global systems integrators” and taking enterprise AI “from pilot to live, fast” through a named, fixed-shape discovery sprint. It validates two of our own moves directly — the anti-incumbent-SI positioning, and productising the entry assessment as a packaged offer rather than a bespoke scope. But it diverges where it counts: horizontal across seven sectors rather than niche-deep, committed to a single US cloud stack rather than sovereignty-aware and multi-model, and selling to enterprise and public-sector buyers rather than formation-stage entities. It is the closest analogue to our delivery shape — and the clearest picture of the real medium-term threat (below): not a Big 4 awakening, but a nimble firm reinventing itself in our direction.

Each proves a load-bearing element of our model in market. None occupies our position: embedded digital operations for formation-stage companies in a regulated, capital-intensive vertical.

The honest gap and the real defence

The competitive census in our specific niche is incomplete: small AU energy-tech advisory boutiques may partially overlap, and verifying that field is part of the go-to-market research sprint already underway — we’d rather find them on a map than in a deal. The medium-term threat isn’t the Big 4; it’s a copy of us, founded eighteen months behind by people who read the same signals — and the Robots & Pencils pattern shows that reinvention is already happening one segment up. The question is whether any of them come down-market into the regulated, formation-stage niche before our sequence compounds.

The defence is sequence and compounding. First-mover in a referral-dense investor community sets the reference point; the regulatory knowledge moat (AEMO, NEM, CER) deepens with every engagement; and the day-one architect position converts into switching costs no late entrant can offer against. The plan’s risk section says it plainly: the moat at day zero is the founding team and its network; the moat at year three is everything that compounded in between. The race is to year three.

09. Financial model: the practice P&L

This is a worked operating model at three scale points — five, ten, and twenty clients — built on the base-case retainer of $15,000 per client per month (section 9). All figures AUD, salaries fully loaded (superannuation plus on-costs), and every load-bearing assumption is named and stress-tested below. The model’s defining property: revenue scales with clients, the cost base scales with senior headcount, and the production layer in between costs almost nothing — so margin expands with scale instead of compressing the way it does in a pyramid.

The P&L at three scale points

Three structural observations. First, the practice is profitable at five clients with three people — no external capital is required to reach viability. Second, margin expands with scale because adding clients adds compute and senior attention, not a junior bench. Third, the AI and platform line — the entire production layer — is roughly 1% of revenue at every scale point. In a traditional firm the production layer is the cost base.

What it takes to serve twenty clients

Cost assumptions

Fully-loaded annual costs, benchmarked against current Australian market data: Client Leader (Director/MD calibre) $280K–$380K, base case $300K; Engagement Architect (senior manager calibre) $165K–$220K, base case $185K; AI Facilitator $85K–$115K, base case $100K. The AI and platform line covers the full model stack — both one-off production and the persistent operate layer (the intelligence layer, monitoring, ongoing agents) — plus retrieval infrastructure, workflow tooling and SaaS: $6,600 to $18,000 a year at five clients, modelled conservatively at $12,000 and scaling sub-linearly.

What keeps that line at roughly 1% of revenue even with an always-on operate layer is architecture, not luck. We run a tiered model stack — on-prem/local models that are effectively free beyond infrastructure, open-weight models via API for routine volume, and frontier models reserved for the hardest judgment-grade calls — with routing deciding which tier handles what. One of the founders already runs exactly this hybrid in production (a locally-deployed ~27-billion-parameter open model alongside a frontier model, in a live quant system), so the cost-management approach is proven, not theoretical — and the local tier doubles as the model-portability hedge behind our sovereignty position. For calibration: a single junior analyst’s 40 hours of research costs a client $6,000–$12,000 through a traditional firm; the equivalent AI production run costs tens of dollars.

Pricing scenarios at five clients

The breakeven retainer with the full three-person team is roughly $10,500 per client per month. The first sensitivity below follows directly.

The sensitivities, ranked honestly

1. Retainer rate — existential, not aspirational. Moving from $15K to $10K turns a 30% margin into a loss. Pricing discipline is a survival behaviour: below ~$12K/month the model is sub-economic in any configuration, and section 9 sets the floor accordingly.
2. Clients per Client Leader — the single most load-bearing unvalidated assumption. The model assumes 6–8 embedded clients per senior (versus 2–4 traditionally), on the logic that AI removes the production drag. If the true ceiling is 4, year one needs a second Client Leader and the early profit disappears; if it’s 10+, the economics improve materially. The first engagements are instrumented specifically to measure this.
3. Second-senior quality. At ten clients the second Client Leader is ~30% of the cost base carrying ~half the relationships, with no junior buffer to hide behind. A mis-hire here is the most expensive single mistake available to the practice.
4. Compute cost — production and the persistent operate layer. The running cost is no longer only episodic production; the always-on operate layer consumes inference continuously. We manage it by architecture — routing across a tiered stack (local/on-prem ≈ free, open-weight API for volume, frontier reserved for the hardest calls), which a founder already runs in production. The residual risk is frontier-capability costs rising faster than routing and falling commodity inference can absorb; the ~1% share of revenue gives large headroom before it bites, and the tiered design is the hedge.

Where the model breaks

Stated plainly: retainers below ~$12K/month; failure to hold embedded relationships beyond 12 months; senior mis-hires at both Client Leader seats; or a quality ceiling (section 15) that forces senior time back into production. None of these is hidden in the averages — each maps to a specific test in the first year of operation.

What this model deliberately excludes

Client equity is not in these numbers. Formation-stage engagements in some markets carry small equity components (0.5–1% per engagement is common where it occurs), and a portfolio of such positions across a successful cohort would be a material second return layer. Whether this venture ever takes client equity is a live, deliberately open decision (section 9) with structural and independence implications — so the operating P&L above is built to stand entirely without it. If pursued later, it is upside on top of a profitable practice, never a substitute for an adequate retainer.

10. Pricing and commercial model

The commercial model is a monthly retainer for a running function. No timesheets, no rate cards, no scope-change theatre. The client buys the outcome — their digital and IT function exists, works, and is accountable — and pays for it the way they pay for any other operating function.

The anchor: what the function is worth, not what it costs us

The pricing comparator is not consulting day rates; it’s the cost of the alternative. A formation-stage energy company that built this function internally would hire a CIO/CTO ($350K+ fully loaded), plus delivery and operations capability, plus the 12–18 months it takes to assemble — a seven-figure annual run-rate before its first asset earns revenue, for a function it only needs a fraction of at this stage. Our base retainer of $15K/month ($180K/year) delivers the function — executive judgment included — for half the cost of the one executive hire alone.

We are deliberate about not anchoring to our own cost base. AI gives us a production layer at near-zero cost; that is the source of our margin, not a reason to discount. The market evidence says premium pricing for AI-delivered professional work holds when the outcome is real: Harvey charges ~US$1,200/seat/month into law firms (a ~US$288K annual entry point) and is oversubscribed at an US$11B valuation; Unity Advisory sells senior-only AI-native CFO advisory on outcome-linked fees with US$300M of institutional backing. Clients pay for working functions and accountable judgment. They always have.

The traditional comparator is now anchored in market data (2026-08, → [[sources/sophia-pricing-model-anchors-2026-08-02]]). The like-for-like Big 4 alternative — a Partner and Director part-time over a Manager, Senior Consultant, and Consultant full-time — runs A$1M+ a year at discounted government-panel rates and A$1.7M+ at rack (blended realised rates A$1,200–2,000/consultant-day; M). The base retainer is therefore a 6–9× undercut of traditional embedded delivery, not the 2–3× a light comparison suggests. And the ceiling is proven from the AI side: per-outcome agent vendors (Sierra, Decagon) land A$230–690K annual contracts for a single automated function (M) — six-figure subscriptions for working functions clear the market today.

The tiers

The floor is a governance rule, not a negotiating posture: the financial model breaks below ~$12K, and a practice that discounts its way into a loss-making embedded relationship cannot serve anyone well. Premium pricing applies where the senior layer carries genuinely heavier accountability — critical-infrastructure cyber posture, market-operator interfaces, multiple concurrent asset commissionings.

These numbers are deliberately conservative — priced to land, not to maximise the first deal. This is a land-and-expand model: the entry retainer buys reference clients and proof of the model, while the real pricing power compounds later — the premium tier, value- and outcome-linked pricing as ROI is demonstrated, growth with the client, and the equity upside held open below. Measured against the value delivered — a mission-critical function the client often can’t assemble at any price — there is real headroom by design.

The benchmark risk: the MSP heuristic, not the Big 4

The pricing pressure to plan for is not the expensive comparator; it is the cheap one. Australian managed IT converges at A$140–250/user/month (M, five independent guides), so a 100-person company’s fully managed IT lands at A$170–300K a year: the same band as the base retainer. A fund CFO will benchmark $15K/month against “what does managed IT cost per head” before anything else, and second against a fractional CTO at A$8–15K/month for one to two days a week. The answer to both is the same and should be made before the question is asked: an MSP buys commodity operations (helpdesk, devices, patching) and a fractional CTO buys one person part-time; neither buys an accountable executive function — market-systems and compliance judgment, board presence, and the build-and-run of the digital estate itself. Where a client genuinely needs commodity IT operations too, that is a subcontracted or brokered layer under our function, not a substitute for it.

The entry product: the advisory wedge

“Be our embedded digital function” is a large first ask of a board that has never worked with us. So advisory is the way in — and its productised, repeatable form is a fixed-fee market-entry digital readiness assessment: four to six weeks, a defined artefact (the day-one architecture blueprint — market systems, OT/IT, compliance, ERP, security, sequenced and costed), priced in the $40K–$60K range. It does real advisory work, demonstrates the regulatory-knowledge moat, and produces the document the operate-retainer conversation flows naturally from. This is the tip of the spear: a bounded, low-risk first engagement that earns trust and leads into the recurring function. It is small enough for an infrastructure fund to commission for any portfolio company without committee anxiety, and it is referrable: a fund that buys one assessment can buy five. (Where we already hold a warm, embedded position, the first entry can instead ride a live engagement — see section 12 — but the assessment is the repeatable wedge for every entity after the first.)

How the retainer evolves

Formation stage: full function, base or premium retainer. As the client scales and internal capability grows (section 6), scope deliberately narrows and the retainer steps down with it — we’d rather hold ten right-sized relationships than cling to scope and be managed out. The mature-stage relationship (roadmap, major decisions, the standing intelligence layer) is a smaller retainer with very high margin and the durability that comes from having built everything underneath it.

Client equity: deliberately open, deliberately excluded

Whether the venture takes equity or gain-share in clients is one of our named open decisions. The case for: formation-stage companies conserve cash, equity aligns incentives, and the eventual upside on a successful energy portfolio is material. The case against, which currently wins: much of our beachhead cohort is fund-owned SPVs and consortium entities where provider equity is unusual; equity introduces independence and structural complications that interact with the vehicle question; and a model that needs equity to work is a model with an inadequate retainer. So the position is: fee and retainer first, the practice profitable on fees alone (section 8), and the equity question revisited deliberately — with eyes open and the right structure — once the practice has standing. What would have to be true: a chosen structural vehicle that accommodates it, a client segment where it’s culturally normal, and a fee base that doesn’t depend on it.

11. Sector selection as the repeatable engine

The biggest failure mode available to this venture is entering the wrong sector — a niche without a real cohort that can’t build the function itself, or one an incumbent already serves adequately. So sector selection is governed, not vibes-based: a scored rubric, a stage-gate from candidate to beachhead, and a written evaluation for every sector that reaches serious consideration. The discipline matters twice over — it protects the first bet, and it is the repeatable engine if the venture becomes a playbook across multiple niches rather than one company in one sector.

The rubric

Five gating criteria, all of which must clear a high bar:

1. A cohort that can’t build the function itself exists now — organisations under heavy regulatory load that can’t staff or afford it (formation-stage companies are the sharpest case, not the only one), not a hypothetical market
2. High digital complexity from day one — real systems required to operate at all: market interfaces, OT, compliance, data
3. The production layer is AI-addressable — the work is research, analysis, documentation, monitoring, compliance
4. Incumbents structurally can’t serve it — a genuinely unoccupied position, not just an underserved one
5. Senior judgment is genuinely required — the client needs accountable executive decisions, which defends against pure-software competitors

Six supporting criteria shape and de-risk the play: a referral-dense backer community; a regulatory knowledge moat; equity/gain-share cultural norms; founder network advantage; policy and capital tailwinds; compounding switching costs. And a list of disqualifiers any one of which parks a sector: no real cohort, physical work AI can’t carry, a capable incumbent in place, prohibitive trust barriers.

The worked example: AU renewables, evaluated and ratified

The framework has now run end-to-end. Australian renewable energy was evaluated in June 2026 and ratified as the beachhead. Each criterion is judged by conviction and evidence, not a self-assigned mark:

• Locked — structurally true; a condition of operating at all
• Strong — a strong case, with the supporting census still underway
• Gap — a flagged risk we design around, not a dealbreaker

Gating criteria — all must clear the bar:

Supporting criteria — shape and de-risk the play:

No disqualifiers. The Locked rows are structural facts rather than hopes — digital complexity and senior accountability are conditions of participating in the National Electricity Market at all, and the regulatory moat and policy tailwind are written into the 2026 reform agenda and the market operator’s system plan. The Strong rows are strong on logic but honest about evidence: the cohort and the competitive field are structurally inferred, not yet census-mapped, and closing that is exactly the go-to-market research now underway rather than something smoothed over. The single Gap — equity culture — is the rubric working as intended: it flagged that much of this cohort is fund-owned, where provider equity is unproven, which is one reason the commercial model holds client equity out of the base case.

The evaluation also produced a sequenced entry: Renewable Energy Zone delivery entities first — the newly formed consortia and network operators standing up REZ infrastructure, where the founding team has direct access — then battery storage operators, then utility-scale generation platforms. Virtual power plant aggregators were parked: they build internal technology teams by identity, which fails the “be the function” premise.

The pipeline behind the beachhead

The stage-gate runs candidate → evaluating → beachhead → active or parked, with every status change logged. Behind AU renewables sit adjacent capital-intensive candidates — data centres (where the founding team’s operator-side history is directly relevant), transport, water — and a standing process for sourcing new candidates as the research layer surfaces them. The discipline cuts both ways: sectors that fail the rubric get parked in writing, with reasons, so the venture’s appetite never quietly overrides its criteria.

12. Go-to-market

Partially scoped — the plan’s most significant open area, named rather than papered over. The strategy now has a shape; the evidence layer underneath it is being built.

What is now decided

• The entry point: Renewable Energy Zone delivery entities (newly formed consortia and network operators), entered through a warm, currently embedded relationship — followed by similar entities across the REZ pipeline. This replaces a cold-start motion with a known first archetype.
• The wedge: the fixed-fee market-entry digital readiness assessment (section 9) — bounded, referrable, and sized for a fund or consortium to commission without committee anxiety.
• The referral thesis: the AU infrastructure-investor community is small and concentrated; one fund relationship is portfolio-wide distribution. This is the structural argument the research sprint must convert into a named map.

The cohort, mapped (research scan, 12 June 2026)

The first-play archetype is not a one-off — it is a production line, and the next units are already visible:

• Live: a recently-closed REZ network operator with a multi-decade operate-and-maintain obligation, construction ramping now, with several gigawatts of generator projects entering the same zone behind it (each one a formation-stage entity with the same day-one obligations)
• Forming now: the New England REZ operator — 6GW across two stages, three consortia shortlisted (anchored by AusNet, Iberdrola, and EDF respectively), commitment deed expected late 2027 and financial close in 2028. All three consortia are prospects today; the winner becomes target entity number two
• Behind that: South West and Illawarra REZ procurements in NSW; five Victorian REZs declared in May 2026 with delivery entities yet to form; Queensland running an incumbent-led model (Powerlink as REZ Delivery Body) where the project SPVs, not the operator, are the formation-stage targets
• The adjacent cohort: 94 grid-forming battery projects in the NEM development pipeline and 24GW of grid-scale storage projected by 2030 — dozens of new operating entities, ratified as the second motion

The scale frame: the market operator’s draft 2026 system plan prices the build-out at roughly $128 billion in annualised grid-scale capital. The venture’s financial model is profitable at five clients with three people (section 8). Against a serviceable cohort plausibly exceeding fifty entities over five years, viability requires single-digit-percentage penetration. The market does not need to be conquered; it needs to yield five clients.

What the research sprint still owes

1. Fund decision-makers and the referral map — the institutional names are known (the two largest local infrastructure managers jointly run $140B+ and a $700M renewable program; a major bank runs a dedicated energy-transition fund); the people and the network overlap are not yet mapped
2. The conversion path — how an embedded contractor relationship and a readiness assessment each convert to a retainer, with the trigger moments named
3. A final competitive sweep — the boutiques found so far sell market advisory and modelling to investors and boards; none found sells an embedded digital-operations function to formation-stage entities. One more pass (managed-service providers) before claiming the gap outright

The honest position on the network (founder session, 12 June 2026)

Beyond the embedded position at the first-play entity, the warm network is unmapped. There are people the founders know without knowing well; there is no pre-existing fund relationship waiting to write a cheque. The founding view, recorded plainly: this is a hustle motion, the venture could fail here, and the team should be under no illusions. The network-mapping exercise below is therefore a genuinely three-founder task, not a formality — and the first agenda item once the founding team is in a room together.

Questions for the founding team (next working session)

• Map the warm network — collectively: which funds, developers, and consortium principals do the three founders actually know today, and at what strength?
• The first-ten list: which ten entities would we approach first if approval/exit resolved tomorrow? (Screened against employment restraints before any approach.)
• What does the embedded-contractor entry at the first-play entity realistically convert into, and on what timeline?
• Who fronts the GTM motion publicly while employment constraints are live — and what can be prepared without crossing the line?

13. The first engagement, scoped

Drafted from the founder working session, 12 June 2026. The engagement design below reflects the embedded partner’s ground truth and the founding team’s sequencing intent. Items still open are listed at the end.

What the first-play entity looks like from the inside

The pattern is familiar from the founding team’s prior embedded work in billion-dollar infrastructure businesses: organisations with heavy obligations and genuinely low digital maturity, because technology is not what they do. These are construction and utility organisations at heart. They carry plenty of hard, hardware-based technology, which brings its own software problems, but the corporate digital layer is thin: no real internal team, most capability outsourced, applications bought piecemeal that don’t mesh, no architecture or strategy holding it together, and procurement capability that is weak precisely where their regulatory exposure makes procurement matter most.

That last point is the live opening. The embedded founder is currently supporting the entity through a system procurement (governance/risk/compliance class), built the entity’s budget in his current engagement, and has developed a proof-of-concept tool on his own initiative. The entity hasn’t seen the tool yet — that demonstration is a card the founding team holds, to be played when the first bounded engagement is framed.

The delivery thesis this engagement must prove

A regulated utility cannot let its engineering team, let alone its finance function, build mission systems with AI directly, no matter how capable the models get. The accountability does not exist: when something fails, the organisation needs a counterparty, support, and a throat to choke, and the frontier AI companies are not providing that layer. The venture’s position is that the AI does the production work, but what the client buys is a supported product with a responsible party behind it. Engagement one exists to prove that position holds in regulated reality.

The architecture follows from the same logic. This cohort already self-manages data centres because regulation demands it, so delivery is designed local-first: a tiered model stack — on-prem/local models for the regulated core and routine volume, open-weight models via API, and frontier models reserved for the hardest judgment-grade calls — with routing deciding what runs where. That design does triple duty: it holds compute cost down (section 8), it keeps sensitive data inside the client’s own environment, and the local tier is the portability hedge behind our sovereignty position — if a frontier model is recalled or restricted (as one was, worldwide, in June 2026), the regulated core keeps running. This is not theoretical: one of the founders already runs exactly this hybrid — a locally-deployed open model alongside a frontier model — in a live, regulated-grade quant system.

Scope and entry

• Entry: the live procurement, not a cold pitch. The work the entity already trusts the partner with extends naturally into the first bounded deliverable; the productised readiness assessment (section 9) remains the repeatable wedge for entities two onward.
• Bounded first, function later: one deliverable with a defined edge, instrumented from day one (senior hours consumed, correction rates), not “be the digital function” from day zero.

Sequencing — settled in principle

Two of the three founders proceed now and prove the model; the third remains entirely outside the venture until a clean exit from current employment, then joins. Nothing runs through the employed founder in the interim. This is deliberately a trust-based structure: if the founding relationships cannot carry that arrangement, the venture was never going to work. (Subject to legal review before any client contract is signed.)

What success looks like — the 90-day bar

One signed, paying, retained client within 90 days of launch. That is the whole threshold. The benchmark that sets it: new tier-two-and-below consulting firms routinely take far longer than a quarter to land their first client, and some fail without ever landing one — so a signed and retained client inside 90 days, entered warm, would put this venture meaningfully ahead of the reference class. Everything else is instrumentation rather than success criteria: senior hours consumed per month and correction rates (the quality-ceiling evidence), and whether the engagement generates a referral conversation (the GTM evidence). A second and third client then prove the revenue model repeats, and that proof is the scaling story: demonstrated reliable revenue in a regulated cohort, ready to run across the REZ pipeline.

Still open

• Who inside the entity sponsors the first contracted engagement, and at what budget line
• Where the boundary sits between the partner’s current personal contracting and the venture’s first contract — and what the legal review says about the transition

14. Risks: what would make this fail

A plan that can’t name its own failure modes isn’t ready to be shown to anyone. These are ours, in roughly the order they could kill the venture, with what we’re doing about each and — where the honest answer is “we don’t know yet” — the specific test that will tell us.

1. Go-to-market is unproven

The risk: Everything in this plan downstream of “win a client” is modelled; winning the client is not. Nobody on the founding team has yet sold an embedded engagement to an infrastructure fund’s investment committee, and the referral-density argument — one fund relationship opens a portfolio — is a structural hypothesis, not a track record. What we’re doing: A dedicated research sprint (fund landscape, named cohort census, competitive field) before any outreach; a deliberately bounded entry product (the fixed-fee readiness assessment, section 9) sized so a fund can say yes without committee anxiety; and a first play that starts from a warm, embedded position rather than a cold pitch. The market test: Does the first readiness assessment convert to a retainer, and does the first retainer generate a referral.

2. The trust barrier on critical infrastructure

The risk: We’re asking boards to let a new three-person firm run the digital function of essential-services infrastructure. However good the model, an investment committee can rationally prefer a worse answer from a bigger brand. What we’re doing: Entering through bounded, low-stakes-first engagements; leading with the founding team’s infrastructure delivery record (the embedded data-centre-operator function, two decades of mission-critical government delivery); and letting the regulatory-knowledge moat show in the first artefact rather than asserting it. The market test: The first engagement won without a brand behind us.

3. The clients-per-senior ceiling

The risk: The economics assume a Client Leader can carry 6–8 embedded clients because AI removes production drag. If the real ceiling is 4 — because relationships, board rhythms, and accountability consume the freed time — year-one profit disappears and the model needs repricing (section 8 shows the maths). What we’re doing: Instrumenting the first engagements to measure where senior time actually goes; capping intake until the number is known rather than discovering it through service failure. The market test: Measured senior hours per client per month across the first 3–5 engagements.

4. The quality ceiling — and the trust bar on the intelligence layer

The risk: Two related forms. At some engagement complexity, AI production may need materially more senior correction — eroding both margin and speed; we don’t know where that ceiling sits for embedded operations work in a regulated market. The persistent operate layer raises a sharper version: an embedded intelligence layer that answers a CFO or board in plain language must be right — a confident wrong number is worse than no number, and decisions get made on it. What we’re doing: Tracking correction rates and rework by task type from engagement one; designing deliverables and the intelligence layer so the highest-stakes outputs carry senior review and verifiable sourcing regardless of AI confidence. What we sell is reliability a regulated organisation can stand behind, not raw answers — which is exactly the standard a bought tool or a DIY chatbot doesn’t reach. The market test: Correction-rate data from the first engagement, reviewed quarterly; and no unreviewed high-stakes output reaching a decision-maker.

5. Sector misjudgement

The risk: The beachhead could be wrong — a thinner formation cohort than the structural evidence implies, or a niche incumbent we haven’t found yet. What we’re doing: This is the risk the sector stage-gate exists for. The beachhead passed a scored evaluation rather than being assumed (section 10), the evaluation’s own evidence gaps define the current research, and the pipeline keeps adjacent sectors warm so parking the beachhead would be painful but not fatal. The market test: The cohort census — if the named-company list is short, we want to know before the first hire, not after.

6. Founder structure and alignment

The risk: Three founders misaligned on model, commitment, or equity is an early-stage killer everywhere; in a senior-judgment business the founding team is the asset, so the risk concentrates. The structural-vehicle question (single entity, backed, per-sector) is also unresolved, and several commercial options hang off it. What we’re doing: Roles, equity, vesting and step-back scenarios are being worked explicitly with the founding team before launch (section 14), not improvised after; the vehicle decision has a deliberate deadline tied to the first engagement rather than drifting. The test: A signed founders’ agreement that all three would re-sign after reading the worst-case scenarios.

Two risks we’re often asked about and deliberately rank lower: AI cost inflation (the production and operate layer is ~1% of revenue, held there by tiered model routing — costs could rise many-fold before the model noticed) and incumbent response (their structural trap is the thesis; the real competitive risk is a fast follower copying us, which is a race we can run — section 7).

15. Funding and the ask

Posture settled at the founder working session, 12 June 2026; numbers still to be done.

The posture

The venture bootstraps. Client revenue funds operations from the first engagement, topped up by what the three founders can put in for equipment and set-up; the financial model (section 8) shows the practice profitable at small scale with three people, so external capital is not required to reach viability. That is a direct consequence of scaling on software rather than headcount — there is no junior bench to fund — and the early discipline it imposes (revenue or nothing) is the same proof an eventual investor would want anyway.

Why external capital still comes, eventually

The founding view is that scaling beyond the first sector will take external money — and not primarily for the cash. The right capital partner brings the things bootstrapping cannot: relationships into the infrastructure-investor community the venture sells to, the credibility that institutional backing confers on a young firm selling to regulated boards, and people who can point the venture at doors it doesn’t know exist. The Unity Advisory precedent (US$300M from Warburg Pincus) shows institutional capital actively backs this model class. The honest caveat: the founding team has not raised before, and this view is an assumption to be tested — against advice, and against the structural-vehicle decision (single entity versus backed platform) when it is made.

Sequencing

1. Now → first clients: founder-funded set-up, client revenue, no raise
2. Revenue proof (one to three clients, repeatable model): the fundraising story exists — demonstrated revenue in a regulated cohort, a playbook ready to run into the next sector
3. Raise (optional, structural-vehicle dependent): capital bought for scale, relationships and legitimacy, on the strength of evidence rather than projections

Still open

• How much runway stages 2–3 actually need per founder (links to section 13)
• What Unity-style backing would cost in equity and independence, versus staying single-entity
• Who the right capital partners would even be — part of the fund-landscape research underway