For fifteen years, SaaS ran on one rule: acquire a customer cheaply enough that their subscription eventually earns more than it cost to win them. Everything else, the free trials, the self-service signups, the land-and-expand motion, was built to protect that ratio. It worked so well that an entire generation of go-to-market teams stopped questioning it.
Then AI arrived, and the ratio stopped being the point.
What actually made SaaS unique
Strip away the funding rounds and the acronyms, and SaaS was built on a handful of structural bets that older software models never had to make. Efficiency was the first. Unlike the old model, where a vendor collected most of a deal’s value upfront, a SaaS company only breaks even on a customer months into the relationship, so every dollar spent acquiring them has to be recovered slowly. That single constraint shaped everything downstream, from freemium tiers to the entire self-service category.
The second bet was distribution. SaaS was global from day one. A ten-person startup in Rotterdam could sell to Sydney by lunchtime, but so could every competitor selling into Rotterdam. And the third was a shift in who does the convincing. SaaS moved buying decisions away from a handful of rainmaker reps and toward a self-service motion where the product, not the salesperson, does the qualifying. Buyers try before they commit. The risk moved from the buyer’s side of the table to the vendor’s.
Where the AI playbook breaks
Sell AI the way you sold SaaS, and you hit trouble fast. I’ve watched this happen repeatedly in the sales calls I review for a living.
The first break is value framing. SaaS pitched optimization: do the thing you already do, faster. AI, especially agentic AI, pitches redesign: don’t do the thing at all, let the system own the outcome. Those are different conversations that need different proof.
The second is the buyer. A SaaS deal usually gets qualified by the end user and rubber-stamped by their manager. An AI deal that actually replaces a workflow gets pulled straight to the CFO or the COO, because you’re not asking someone to adopt a tool, you’re asking them to retire a process. That matches what BCG’s own research has been confirming: the AI buying center sits higher in the org chart than the SaaS one ever did.
The third break is proof. SaaS sold usability: can your team pick this up in a week? AI has to sell a head-to-head, human-versus-agent comparison, with a delta the buyer can defend to their own board. And the fourth is pricing. Seat-based pricing assumes you’re selling access to more humans. When an agent does the work instead of assisting a human who does it, charging per seat is charging for something the customer no longer needs.
If AI is replacing a process rather than assisting a person, selling it like SaaS means selling the wrong unit of value.
The closest working analogy isn’t software at all. It’s how an automotive OEM works with a parts supplier: deep technical credibility, long procurement cycles, high switching costs, and a relationship built on dependency rather than a quick trial. That’s closer to how enterprise AI actually gets bought than any SaaS deal-cycle benchmark.
The obstacle nobody puts in the deck
Here’s the part vendors consistently underestimate: the biggest resistance to AI adoption rarely comes from the C-suite that approved the budget. It comes from the middle managers whose teams the AI is meant to shrink.
Going from managing fifteen direct reports to five isn’t how most corporate cultures reward efficiency. It’s how they read a demotion. So when adoption is left to bubble up organically, team by team, middle management has every incentive to slow-walk it, and uncontrolled, siloed automation layered on an already broken process just cements the mistakes at greater speed. I’ve written about this pattern in more detail on my blog, and the fix is structural, not motivational: AI adoption has to be a board-level mandate, not a bottom-up initiative that depends on the same layer of the organization it’s about to disrupt.
What this means for the pitch
None of this makes SaaS obsolete. It makes the SaaS sales motion the wrong template for an AI deal. The rep profile that wins here looks more like a consultant who understands the customer’s process than an account executive who understands the customer’s dashboard, and the qualification conversation has to move up the org chart before it can move back down into the workflow.
Most of the sales teams I work with inside my sales coaching program are still running SaaS-era discovery on AI-era deals, and it shows up in the numbers before it shows up anywhere else. Fix the framing first. The pricing model will follow.
Nils Brosch is a B2B SaaS and AI sales trainer and consultant working with teams across Benelux and DACH. He has personally reviewed more than 1,400 sales calls across 140+ companies. More at nilsbrosch.com.
