A podcast argument from a finance AI vendor separates two things the industry keeps conflating: the model layer and the venue layer — and the structural logic holds even after you strip out the pitch.
Disclosure: This article summarises a podcast conversation. The central argument belongs entirely to Gabe Stengel, CEO of Rogo, a company that sells AI software into financial services — meaning the thesis is also his company’s commercial positioning. This is not investment advice.
On Invest Like the Best, Gabe Stengel — CEO of Rogo, which sells AI software into finance — made an argument about where value settles in AI-era capital markets. The podcast titles the episode “Why OpenAI and Anthropic Won’t Win Finance,” but that framing is the show’s. Nothing in the argument, or here, asserts that any laboratory will or will not win any market.
Stengel’s argument turns on a distinction between intelligence as a commodity input and the venue where consequential actions are recorded. His example is a compliant, secure data room for a public-company acquisition — something integrated into the way work is actually done rather than a static folder. His claim that the labs would never want to build a data-room business is speculation about their intentions, not a statement of their plans, and no laboratory has said so here.
The interest in the argument is structural rather than predictive. The same shape appeared independently in agentic commerce that same week: wherever an agent must transact rather than merely answer, the binding constraint moved off the model and onto the permissioned venue where the action is recorded. Two industries, the same week, the same underlying property.
The key insight: Intelligence can be sold to every participant in a market simultaneously. A regulated, compliant transaction venue cannot — it carries custody obligations, sector-specific liability, and switching costs that belong to an entirely different businessmodel. These are not adjacent; they are separate undertakings that happen to share a customer.
The Structural Read
The Map of AI framework distinguishes layers by what defends them. Model capability is defended by research investment and compute. Venues — systems of record, transaction rails, communication infrastructure — are defended by compliance, custody, integration depth, and the cost of moving data out. Those are different moats, built by different capabilities, regulated by different bodies, and attracting different margin profiles.
Stengel’s structural point, separated from his forecast, is that a horizontal supplier taking on vertical custody accepts a different regulator, a different failure mode, and a materially different cost base. That is the ordinary reason horizontal platforms decline vertical surfaces — not preference, but incompatibility of obligations. His claim about what the labs want is weaker framed as psychology; it is considerably stronger framed as structural incompatibility, which is the version worth keeping.
The second half of his argument is about tacit knowledge. The move he describes — converting expertise that lives in people into an owned, auditable system — is a real and separate value-creation mechanism. Expertise that walks out with an employee cannot be sold with the company, cannot be audited, and does not scale beyond working hours. The tension is obvious: that project is precisely what Rogo sells, which is why it should be weighed as an argument from an interested party.
Gabe Stengel — Invest Like the Best
“If you actually want to be the exchange for all of high finance and all of capital markets, you not just need to own the intelligence, you need to own the transaction venue, the communication venue, the workflows, and all the data inputs that go into it.”
Map of AI — Layer Logic
Answering is unconstrained. Transacting is not.
An answer can be generated anywhere, by any model, with no record that satisfies a third party. A transaction happens somewhere specific, under someone’s rules, with a record that must survive an audit. The binding constraint in any agentic workflow is not the intelligence layer — it is the permissioned venue where the action is logged.
Three Implications
IMPLICATION 1 — The Venue Is a Separate Asset Class
Systems of record in regulated industries — custody rails, compliant data rooms, audit-ready communication logs — are not model-layer problems. They are defended by switching costs, regulatory approval, and integration depth. A better model does not dissolve those barriers; it runs on top of them or is excluded by them.
IMPLICATION 2 — Horizontal Suppliers Face a Structural Ceiling in Vertical Markets
The reason a high-margin horizontal platform does not simply absorb a vertical surface is not reluctance — it is that doing so changes the regulatory relationship, the liability exposure, and the coststructure in ways that compress the original margin profile. The ceiling is structural, not motivational. This applies whether the supplier is an AI laboratory, a cloud provider, or a software platform.
IMPLICATION 3 — Tacit Knowledge Conversion Is the Overlooked Moat
Stengel’s second argument — encoding expertise from people into owned, transferable systems — is analytically separate from the venue argument and survives independently. Expertise that lives only in individuals depreciates with attrition and does not appear on a balance sheet. Expertise encoded into an auditable system is a different kind of asset. The catch: the companies best positioned to facilitate that conversion are also the ones selling the encoding infrastructure.
Business Engineer Framework
The Map of AI — Where Value Actually Accumulates
The Map of AI maps more than 200 companies across nine distinct layers — from compute and model infrastructure through to the venue and workflow layers where Stengel’s argument says the durable value sits. Understanding which layer a company occupies, and what defends it, is the first step to evaluating arguments like this one on their structural merits rather than their pitch packaging.
Strip the forecast and the vendor framing from Stengel’s argument and what remains is a precise structural observation: intelligence is an input that can be sold everywhere at once, while the venue where consequential actions are recorded is a different asset entirely — one defended by compliance, custody, and switching costs that no model improvement dissolves. Whether he is right about who wins is unknowable and unpredictable; whether the distinction between the model layer and the venue layer is real is not a matter of prediction at all. It is already visible, simultaneously, in capital markets and in agentic commerce. The shape is the thing worth holding onto.
Structural reads like this one, every week — no noise, no forecast cosplay.
This summarises a podcast conversation, not data, and it is not investment advice.Gabe Stengel runs Rogo, which sells AI software into finance — the thesis that value accrues to workflow and venue rather than to model providers is therefore his company’s own positioning, and should be weighed as an argument from an interested party. The episode title is the show’s framing; nothing above asserts that any laboratory will or will not win any market, and nothing above makes any claim about any laboratory’s plans — his remark about what they would want to build is speculation about their intentions. The figure of 90% of enterprisevalue in ten years is his forecast, not a measurement, and is not endorsed here; the ten thousand agents he describes are a hypothetical he poses, and his comment that models are smarter than anyone he knows is his opinion about his own acquaintances rather than evidence about model capability. No Rogo revenue, customer, funding or valuation figure appears above, no market-size or adoption figure, and no bank, fund or customer is named.
Gennaro is the creator of FourWeekMBA, which reached about four million business people, comprising C-level executives, investors, analysts, product managers, and aspiring digital entrepreneurs in 2022 alone | He is also Director of Sales for a high-tech scaleup in the AI Industry | In 2012, Gennaro earned an International MBA with emphasis on Corporate Finance and Business Strategy.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
