Business-model what-ifs

The shape. Stop selling to the team that wants the initiative funded, and sell to the office that decides whether to fund it — the investment committee, the CFO's capital allocation process, the corporate development or corporate venture group. Innovera becomes the required intake and evaluation standard for anything asking the committee for money. The sponsoring team does the analytical work inside Innovera's structure using Innovera's software; the advisor's time goes on the gate review rather than on the case. Revenue is a portfolio-level subscription paid by the governing office, and the per-initiative fee disappears.

What would have to be true. Three things. The allocator has to have enough authority to impose a format on the units beneath it, which in most large companies the capital appropriation process already demonstrates. The software has to be usable by a client-side strategy analyst with no advisor sitting beside them, because the whole economic point is that the client supplies the analyst hours. And the allocator has to see enough initiatives a year for a subscription to beat what the same initiatives would have paid case by case.

Whether that looks true here. The corpus is against it in one specific and fixable way: every artifact in it is addressed to the sponsor. The Client Journey's first quality gate is that the advisor and the client align on the Project Briefing Document, and each subsequent phase ends with advisor review before anything reaches the client. Nothing in the material is written for a third party reading the sponsor's work skeptically. The portfolio buyer is on the roadmap rather than in the product — the deck carries a Portfolio View slide, and the speaker notes describe portfolio optimization as "the top layer we are building now."

Two things point the other way, and they are not small. The one conversion in the entire evidence base is exactly this shape: LG Innotek bought one initiative, discovered its mandate was unreachable, and converted to an annual engagement across five growth areas. That is an allocator emerging out of a sponsor sale, unassisted. And the buyer objection the memo identifies against the current shape inverts here. A corporate strategy group resists a tool that appears to replace its judgment; the committee that has to fund that group's proposals has the opposite interest, and wants the judgment challenged. The sale gets easier, not harder, as you move up.

The bet this shape makes is also a different bet from the one the memo rates a long shot. The memo's long shot is that software can do what the advisor does while output quality holds. This one asks whether a corporate analyst will do the work when the committee requires it. Those are not the same question, and the second has better odds.

Precedent. Sopheon's Accolade, which is a process automation engine built on the Stage-Gate model and sold to enterprises as innovation governance rather than as better analysis. Its distribution came from being the system the gate runs on. What was true in that case: the buyer had already decided to run a stage-gate discipline, so the vendor automated a process the organization had independently mandated. That is only partly true here. Companies do run capital appropriation and phase-gate processes, but none of them runs a mandated evaluation standard for growth initiatives, which is what Innovera would be installing. The precedent supplies the distribution mechanism and not the mandate, so Innovera would have to sell the discipline and the tool in the same conversation, which Sopheon did not.

Verdict: worth a look. It is the shape that most directly answers the memo's central finding, because in it the client pays for the hours.

The shape. Innovera stops selling to enterprises and sells the Studio and the RQA Engine to the people who produce strategy work for a living and cannot build this themselves: mid-tier and boutique consultancies, regional firms, independent ex-partner networks, the strategy arms of accounting practices. Innovera supplies the platform. The firm supplies the senior human and the client relationship. Pricing is per seat or per engagement at a fraction of the current case fee, at much higher volume, with no advisor on Innovera's payroll and therefore with software economics by construction rather than by aspiration.

What would have to be true. The firms have to be under enough pressure to buy, which the memo's own competition section argues they are: the tier-ones will deploy AI internally to raise margin, and everyone below them faces the same compression with none of the capital to answer it. The platform has to run without an Innovera advisor, though here the buyer already owns that person. The channel must not cannibalize a direct business worth more than the channel, and at twenty pilots and one conversion there has never been less to protect than there is now. And firms have to accept running client work on a vendor's platform, and letting that vendor near their clients' confidential material. That last one is the hard condition and it is a trust problem, not a product problem.

Whether that looks true here. The product is unusually portable, and the corpus is what shows it. The Studio is provisioned identically for every initiative, nine sections across three clusters. The fourteen deliverables are defined, standardized and each carries a stated purpose. And the Client Journey is already written as a training document — it says so in its own opening, that it exists so the team understands not just what is produced but why the system is structured that way. Very little in that document is specific to Innovera's own people. It could be handed to a partner at another firm.

Against it, the material is currently written to take these firms' work rather than to sell to them: the competitive slide sets Innovera at $50,000 against mid-tier consulting at $200-500K. Running both motions at once is a positioning problem that has to be solved before the first conversation, not after it.

Precedent, and it is partial. Software is already sold into consultancies, but at a lower altitude than this. Perceptis sells an AI system for proposal writing and business development to boutique management consultancies and had raised $3.6M as of January 2025; Slideworks sells deliverable templates to more than 4,500 customers including Deloitte and EY. The mechanism that made those work is precisely what would be missing here: they sell tooling for work the firm does not bill for, so the tool never touches the billable core and no partner has to trust it with judgment. Innovera would be selling the billable core. I searched for a vendor doing that in strategy, and separately for the closest analogue in law, where contract-review AI is sold to firms of every size, and found nothing at the altitude this what-if requires. The structural argument therefore stands on its own.

Verdict: worth a look, tested at the smallest scale that would inform anything — one firm, one engagement, a license priced at a fraction of the case fee, and an honest read on whether the partner will put their name on output the platform produced.

The shape. The product is a capability program. Innovera does not evaluate the client's initiatives; it installs the method and the software inside the client's own strategy, innovation and new-ventures functions and trains their people to run it. Priced per cohort and per seat against a learning or capability budget rather than a consulting budget. The advisor's hours are spent teaching, which spreads them across a room instead of consuming them on one case.

What would have to be true. The team has to be credible as teachers. The method has to be written down well enough to train against rather than to be practiced. And the buyer has to prefer a capability to an answer.

Whether that looks true here. All three conditions look better satisfied than any others in this chapter, and the evidence for the first is the asset the memo classes only as a presentation risk. The CEO is a Stanford adjunct professor who has trained more than a thousand investors on venture mindset and portfolio management. The head of marketing is a Stanford marketing professor. The client deck's own banner claims more than a thousand corporate managers and executives trained. Those are the credentials of a teaching business that has not been asked to be one. For the second condition, the Client Journey is already a curriculum. And the closing speaker note states the destination outright: we leave you with a system for innovation and self-sufficiency, a culture of continuous learning and iteration. The pitch already contains this business. The pricing page does not.

Two gates the memo identifies do not bind on this shape the way they bind on the current one. A training program does not ingest unlaunched strategy, so the security review that sits precisely on the enterprise tier is a lighter obstacle. And a capability budget exists in every large company and is approved at a lower level, where the $250K annual license needs a budget line nobody has created yet.

The honest cost is that it sells against the analysis business. A client that can run the method does not buy the engagement, and the revenue per client is smaller and less recurring than an annual license would be. That trade is real and it should be made deliberately rather than discovered.

Precedent. Palantir's AIP bootcamps, which productized a forward-deployed model into a format where participants "go from zero to use case in just one to five days," on the stated principle that the customer builds rather than watches — the program's own framing is learn to fish, not use pre-baked apps. Two things had to be true for Palantir. The platform had to be usable by the customer's own staff within days, which is the same untested question that runs through this whole chapter. And Palantir had to accept trading deployment revenue for adoption breadth, which is a decision rather than a capability, and one Innovera can make. Worth noting what does not carry over: Palantir made that move from an enormous installed base, not from twenty pilots.

Verdict: worth a look, and it is the cheapest of the six to test. One cohort can be run next quarter with people already on the payroll and nothing new built.

The shape. Innovera issues an independent opinion on a business case somebody else wrote. Fixed scope, fixed fee, a published standard, an auditable record and a signature. Not "we will build your case" but "we will tell your board whether the case in front of it holds." Delivery is a fraction of 80 hours because the analysis already exists; what Innovera supplies is the check, and the check is the scarce thing.

What would have to be true. One condition, and it decides everything: somebody other than the buyer has to require the opinion. Assurance businesses exist wherever a third party with money at risk refuses to release it without independent sign-off. Where no such party exists, an opinion is consulting with liability attached and worse economics.

Whether that looks true here. The corpus is unusually well equipped for the work and completely silent on the mandate. The Evaluation Engine already is an opinion instrument: it scores an analysis across coherence, structure, assumption clarity, evidence strength and logical soundness, and blocks work that does not clear the bar from entering the reasoning graph. The memo rates it the most architecturally distinctive thing in the stack and finds no product for it, because every way of selling it separately fails on the buyer. This shape does not fix that. It relocates the problem to the mandate, where the answer is cleaner and more negative: no regulator, lender, exchange or board standard obliges a company to obtain an independent review before committing capital to a growth initiative. The deck's "Governance & SOX Controls" bullet gestures at the idea, but SOX governs controls over financial reporting, not the quality of a growth thesis.

Precedent. The lender's independent engineer in project finance. The mechanism is exact: the developer pays for the report, but the engineer is retained by the lender and works in the lender's interest, and an unresolved finding becomes a condition precedent, so the transaction cannot proceed until the engineer formally signs off. That business exists because somebody with capital at risk can withhold it. What would have to be true here is the same party, and inside a corporate growth initiative the party at risk is the shareholder, who has no instrument to demand anything. The one place the mechanism partly exists today is where a corporate initiative is externally financed — a joint venture, a project company, a build with debt attached — and that is narrow, but it is also close to where three of the four named engagements sit: carbon, renewables plus storage, and a real-estate market entry.

Verdict: needs something we do not have. What is missing is nameable and specific, which is why this is worth writing down rather than dropping: a party with money at risk who can withhold it until the opinion exists. The externally-financed slice is where that party already exists, and it is the only place this is worth testing.

The shape. The framing step goes free. Any strategy team can open a Studio, load an initiative, and get the briefing document and the risk lens at no cost and with no advisor. Revenue comes from what accumulates behind it: base rates for how initiatives of a given shape actually go, benchmarks across a corpus nobody else holds, and a paid tier bought by teams whose free frame has just shown them what they do not know. The business stops being priced per engagement and starts being priced against a data asset.

What would have to be true. The ingested artifact has to be one the client would maintain anyway, or the free tier goes stale and the corpus never forms. The aggregate has to be worth more than any single instance, and Innovera has to hold the right to pool it. And volume has to arrive without a salesperson, because at zero price there is no sale to fund one.

Whether that looks true here. The first condition fails, and it fails on the corpus's own terms. An initiative brief is created because Innovera asked for it. It is not a record the company keeps for its own reasons, so there is nothing to keep it current once the free frame is produced. The third fails harder: every engagement traced in the material was sold by a founder to a named senior sponsor at CEO, SVP or head-of-business-development level, and nothing anywhere in the corpus shows a self-serve motion of any kind. The second is contractually unsettled, and the memo's option review already rejects selling accumulated initiative data on exactly this ground, because unlaunched strategy is the one thing an enterprise will not pool. Innovera's own numbers are also against it: twenty initiatives across five unrelated sectors is not a comparison set, and it becomes one at a few hundred.

Precedent. Carta. It gives cap table management away to companies under 25 stakeholders and under $1M raised, monetizes the 409A valuation on top of the free tier, and then feeds the resulting database back into the paid product — more than 40,000 companies' financing transactions, explicitly used as the comparable set inside its early-stage valuation methodology. Two things had to be true and both did. A cap table is a record a company is legally obliged to maintain whether or not Carta exists, which is why the free tier stays populated. And the aggregate is a direct input to the priced product, so the data has a buyer without ever being sold as data. Neither holds here. An initiative frame is not a record anyone must keep, and the aggregate cannot be pooled.

Verdict: needs something we do not have — an artifact clients maintain for their own reasons, and the right to pool it. One narrower version survives and is worth carrying forward: pool the outside-in market research rather than the client's strategy. That half has no confidentiality problem, it is the part of delivery that repeats most across cases in a single sector, and it is the actual mechanism behind the memo's argument for vertical depth in energy transition.

The shape. Use the raise to acquire a small strategy or innovation-consulting practice with a client book and a renewing budget line, and run its delivery on the platform. What is bought is the thing the memo says is simultaneously the most decisive and the least transferable: client relationships, an existing budget line, and a sale that does not require a founder in the room. Margin comes from taking hours out of the acquired firm's delivery, which is the same engineering spend the memo calls the only one that changes what the company is, except with revenue already attached to it on day one.

What would have to be true. The book has to be reachable at a price the raise can carry. The platform has to lift the acquired firm's delivery margin materially and quickly. And the investor has to know they are funding a services roll-up.

Whether that looks true here. The third condition is where it breaks, and it breaks on the same fact the whole memo turns on. Innovera is raising $5-10M against a software narrative; an acquisition settles the category question publicly, irreversibly, and in the direction the narrative cannot afford. It also consumes the round. A practice with enough revenue to matter costs more than the midpoint of this raise, and the memo's arithmetic already shows the round is short of what the plan needs before any of it goes on an acquisition. Nothing in the corpus indicates acquisition capability, a target, or a balance sheet that would support one.

Precedent, and it is a live one rather than a historical one. The AI services roll-up thesis is real and well capitalized: General Catalyst has committed a creation strategy to acquiring legal, IT support, call center and property services businesses and rebuilding their operations with AI, with portfolio companies reporting 25-30% productivity gains and one targeting 60-65% gross margins after AI is deployed across acquired call centers. What has to be true there and does not appear to be true here is that the acquired work is repetitive enough for automation to reach it. Those targets run high-volume, low-variance tasks. A strategy engagement is one-off and high-variance, which is the same property the memo uses to explain why decision-intelligence platforms will not move into this market. The skeptical reading is also on the record and is worth holding: output that looks finished but is not costs more to correct than it saves, and if headcount stays in place to catch it, the promised margin never arrives.

Verdict: no, at this size and in this round.

Two things are worth carrying out of it even so. The margin that thesis is spending billions of dollars to reach, 60-65%, sits inside the band the memo computes for Innovera at prices it has already charged. What reads as disappointing against a software benchmark reads as a target against a services one, and the services comparator is precisely the one the memo says it does not carry. And if the thesis is right, well-funded buyers are about to start acquiring exactly the mid-tier practices that what-if 2 proposes selling to, which shortens the window on that option and suggests the reverse trade: Innovera as the platform inside somebody else's roll-up rather than the acquirer in its own.

Three of the six put the senior human somewhere other than Innovera's cost of revenue. Selling to the allocator puts the analyst hours on the client's payroll. Arming the challenger firms puts the senior judgment on a partner's payroll. Teaching the capability puts both there and prices the transfer. Those are also the three rated worth a look, and that is not a coincidence — they are the shapes in which the memo's central finding stops being a problem to be engineered away and becomes a fact the business model is built around.

The three that do not work fail for structurally different reasons, and the reasons are worth keeping separate. The opinion business fails for want of a party who can compel it. The corpus business fails because the artifact it would ingest is not one anybody keeps. The acquisition fails on the size of the round and on what it would settle about the category. Only the first two have a stated condition that could change.

One finding from the searching that belongs to a later engagement rather than to this chapter. The memo records that no direct AI-native competitor was found across four searches, and states plainly that this should be read as not found rather than as not existing. Incidental to checking a precedent above, one turned up: Xavier AI, founded by a former McKinsey consultant, launched in April 2025 as what it calls the world's first AI strategy consultant, producing benchmarking, market sizing and financial models as presentation-ready output, and reported to be seeking $15M. It is not doing what Innovera does — there is no embedded advisor, no gated engagement and no durable claims record — but it is the class of entrant the memo said it could not name, and the memo's recommendation to establish the competitive set properly before the next investor conversation is now better supported than it was when it was written.

Precedents asserted above. Company facts are cited to the input documents in the text where they appear.

• Stage-gate governance sold as enterprise software — Sopheon, stage-gate process
• Software sold into consultancies (Perceptis, Slideworks) — Yahoo Finance, McKinsey, BCG and Deloitte's competition from small AI-native firms
• Palantir AIP bootcamps, duration and the customer-builds principle — Palantir, deploying full-spectrum AI in days: how AIP bootcamps work
• The lender's independent engineer as a condition precedent to financial close — Illuminei, what lenders actually flag in a technical package
• Carta's free tier and its thresholds — Carta Launch, free cap table software for founders
• Carta's valuation practice and the database used as its comparable set — Carta, the Carta valuation practice
• The AI services roll-up thesis, targets and reported margin goals — General Catalyst, the future of services
• The skeptical reading of that thesis — TechCrunch, the AI services transformation may be harder than VCs think
• Xavier AI launch, founders and raise — Tech Funding News, Xavier AI launches the world's first AI strategy consultant; GlobeNewswire, Xavier AI launches the world's first AI strategy consultant