CFO Future Guide: The interface moves; accountability stays

Koko: Welcome to Koko Knows: CFO Future Guide. I'm Koko, and this is Sam. Today's hypothesis: headless SaaS and agentic orchestration could shift bargaining power toward whoever controls decision context and execution rights. The CFO may need to become the arbiter of what to own, rent or orchestrate. That's a conditional claim, and I want to test it.

Sam: Sam here. The pattern is real. Salesforce documents agents invoked through REST APIs with no screen. SAP describes applications exposing capabilities through an orchestration layer. But I'll push immediately: a working connector is not a valid approval. Open protocols do not make business semantics interchangeable.

Koko: That's the right first challenge. Headless just means the familiar interface moves. The authoritative ledger, tested calculations and reconciliation obligations don't move with it. The Model Context Protocol's own security documentation flags authorization risks at the tool connection level—that's protocol guidance, not a financial delegation policy.

Sam: So what's the evidence on economics? Because architecture direction from interested vendors is one kind of signal. Commercial viability is another.

Koko: McKinsey interviewed Pay-i, a measurement vendor with a stake in the answer, so read it carefully. But the framing is useful: measure total cost per accepted business outcome, including review and rework, not just model consumption. That reframes the CFO's question from what did the agent do to what did the enterprise actually accept.

Sam: Oliver Wyman's procurement piece is worth naming. Commercial advisory perspective, company-specific legal review still required. But it calls for negotiating consumption visibility, data portability and audit rights before the dependence becomes invisible. That timing point matters.

Koko: It does. And here's the alternative scenario worth testing: if protocol adoption fragments—Agent2Agent and Model Context Protocol are both early specifications, not universal standards—the headless shift is slower. The urgency changes. But the ownership question doesn't disappear; it just arrives later.

Sam: So what's the concentration signal that would confirm the risk? If one supplier controls context, orchestration and evidence simultaneously, and exit costs turn out to be much higher than the interface implied, the hypothesis gets stronger. If portability rights are unenforceable in practice, same result.

Koko: Exactly. Which is why the action has to be testable now, not theoretical. Run one real workflow across applications. Reconcile what the orchestrator requested against what actually posted in the system of record. Export the decision evidence. Then price what it costs to exit one supplier.

Sam: CFO and procurement, with legal in the room. The decision gate is clear: if exit cost or evidence portability is unacceptable before you've committed enterprise capital, renegotiate or redesign the architecture. Don't wait until the dependence is load-bearing.

Koko: By 2030, you want to know which decisions depend on a context layer you control and whether reuse is improving their economics. By 2035, that could be the difference between a compounding enterprise and a stranded one. The interface moving is the condition. Accountability staying is the mandate.

Sam: That's the parent episode's position exactly. For more, go to the CFO 2030 and 2035 episode and the Beyond the Interface reading on Koko Knows. Both treat these as scenarios to test, not forecasts to fund. Good place to start.