Originally published at Forbes
Software markets have had a wobble. Public SaaS valuations have softened, growth multiples have compressed and the phrase “SaaSpocalypse” has crept into investor vocabulary. The catalyst is clear: agentic AI.
In July 2025, I wrote about the evolution of agentic AI in payments. Fast-forward eight months, and Mastercard and Visa have just announced agentic payment pilots. And suddenly, the spotlight is back on CEOs—I’ve observed in some cases that those who don’t mention agentic road maps are causing investor nervousness.
The selloff was extreme—$285 billion wiped off SaaS share prices—but interestingly fairly uniform. This is a true market error, as agentic AI will not impact all SaaS companies equally. What it will reveal is which companies are truly AI-first and which have bolted on AI to their existing product set to appease the investment world. And there will be nowhere to hide.
The Worker-Agentic Shift
The latest well-reported narrative is of Anthropic’s release of an agentic tool, Cowork, in combination with markdown files explicitly emulating sector-specific specialists (lawyers, accountants, sales, marketing, etc.).
We know the immediate future, at least as it’s often described: Agentic AI turns AI from helpful assistants to executors. They plan multistep workflows, interact with other software, apply rules and escalate only when necessary. They begin to resemble digital colleagues rather than digital tools.
In payments, imagine intelligent agents capable of dynamically adjusting fraud controls in real time, rerouting transactions to optimize cost, managing cross-border liquidity positions or resolving disputes within defined compliance guardrails. That is not incremental automation. It is operational redesign.
The impact ripples outward.
Contagion And The Finance Sector
There have been murmurs online of the potential for contagion. If SaaS is at risk, what if it spread to banking-as-a-service?
This is where it gets interesting; it hasn’t spread, and in my view, it is extremely unlikely to. And the reason why is actually instrumental to the real catalyst of the SaaS selloff.
Rightly or wrongly, the huge drop in SaaS stocks was because of a perceived attack on their revenue base. Most SaaS businesses are priced per seat. That model worked because productivity gains were linked to human users. More employees meant more licenses. More licenses meant predictable recurring revenue.
But what happens when 10 AI agents enable one compliance officer to oversee the work previously handled by five people? What happens when reconciliations, reporting and fraud monitoring are executed autonomously?
The seat stops being a clean proxy for value.
Some software companies have seen this coming and already moved to a consumption model—pay for what you use. But that’s not enough and, ironically, the reason why BaaS is very well insulated.
The future is outcome-based pricing—payment for results, not for effort. And the one area where this is already in place? Banking.
Almost every contract pays on outcomes: the result of a payment processing, not on the attempts to get there. AI may reduce the cost of implementation of the processing, sure, but as a model, it’s already ahead of its time.
The second reason contagion has to be limited is because of the huge infrastructure hurdles new entrants need to overcome to enter the market. Regulatory oversight is mandatory, not a nice-to-have in the banking world, and these barriers to entry slow new entrants into the market.
The third reason is scope. It’s one thing to vibe-code a restaurant booking app in an afternoon, quite another to code up serious infrastructure, payment rails, KYC protocols, permissioning, audit trails, etc. This, of course, didn’t stop enterprise SaaS companies from selling off hard, but I believe we’ll see a level of correction here as the same logic applies. Who wants to vibe-code Oracle in their lunch hour? And not least, the sheer cost of implementation and pain of switching keeps the customer base from experimenting with the new kids on the block.
Where the pressure will really be felt is in narrow, feature-shaped products whose differentiation is largely interface-driven. Not finance, and not even SaaS as a whole. If an intelligent agent can interpret a workflow, call APIs and orchestrate tasks across systems, the stand-alone interface becomes less defensible.
Arthur Mensch, CEO of Mistral, recently remarked that AI is allowing software to be developed “at the speed of light.” When internal tools become dramatically cheaper and faster to build, the long tail of SaaS faces a tougher build-versus-buy calculation.
The Future Of BaaS Versus SaaS
The fact that there was no contagion from SaaS to BaaS is not just telling, it shows a critical difference in how these industries have evolved.
Even though BaaS works with far more constraints than the typical software company, we have collectively not only adopted new technology but are constantly pushing the frontiers of development. Stablecoin payments, crypto exchanges, agentic money-saving experts, shoppers and accountants, to name but a few: all coming to a banking services in the near future. And this has been achieved under the umbrella of intense regulatory scrutiny and the numerous detailed and expensive protocols required for all financial institutions.
SaaS companies? Some enterprise companies have been speedy to adopt and redesign around agentic possibilities; others have simply told their shareholders their AI story and failed to make substantive changes.
The market is collectively punishing the sector for this lack of mobility—in some cases justified and in others overblown. For those companies addressing this U-turn now, the questions to ask are: How deeply embedded is agentic AI in your workflows? How safely can agents work within your environment? Does your pricing reflect outcomes? Are you designing with regulatory constraints in mind from day one?
For BaaS? Almost every single financial institution has been addressing these questions since the advent of AI. There’s no contagion because the very reasons SaaS was punished are the very reasons the fintech sector is still growing.
There’s no room to rest on our laurels, but there is vindication for those who not only saw this coming but acted accordingly.
Comments are closed.