For two years, every AI conversation in B2B SaaS started the same way. How many tools have you rolled out. What percentage of the team is actually using them. How big is next year’s experimentation budget. Adoption was the whole story, and a strong adoption number got treated like an achievement in its own right.
AI for the sake of AI is not a strategy, and it won’t drive the business transformation your customers and your board need. Automating the wrong processes, or investing without a measurable outcome in mind, results in wasted effort at best and an acceleration of the wrong results at worst. This very topic was on the agenda at a recent Primary Venture Partners event, where Jennifer DiRico, CFO at PTC, was one of the speakers. Jennifer wasn’t talking about adoption at all. Her point, as I heard it: executives need to start proving the return on what they’ve already spent, not defending another year of piloting. Nobody in the room pushed back. A year earlier, that same room would have been comparing notes on how many licenses everyone had bought.
That shift matters more than it sounds like it does. “Prove the ROI” is a much harder question than “did we roll it out,” and most GTM organizations were never built to answer it.
The Adoption Era Just Ended
For most of the last two years, moving fast was the whole mandate. Nobody had baselines yet, the tools themselves were changing monthly, and a company that waited to build a measurement system before adopting anything would have been lapped by competitors who just started using the tools. Rolling out quickly was a defensible strategy. Rolling out quickly and never circling back to check the results was not, and that second part is where most companies still are.
You can be fully adopted and still not know if any of it is working. Those are two different questions, and the industry spent two years answering only the first one.
What Ford’s Correction Actually Proves
Around the same time as that event, Ford Motor Company brought back roughly 300 veteran engineers and inspectors it had let go, after the AI tools meant to handle quality checks in their place missed problems only experienced people could catch. Charles Poon, Ford’s VP of Vehicle Hardware Engineering, put it plainly afterward: “Artificial intelligence is a fantastic tool, but it’s only as good as the information you use to train it.” The returning engineers now run quality checks directly and train the tools that were supposed to replace them. Ford went on to take the top spot among mainstream brands in the 2026 JD Power Initial Quality Study, its best result in sixteen years.
The easy read on that story is that AI wasn’t ready. The real lesson is narrower and less comfortable: Ford handed a critical process to AI without ever confirming it could match what the humans doing it were catching. Nobody had proven it worked before the humans were gone. AI didn’t fail here. A company skipped the question that would have told them whether it was working, and the results asked it for them instead.
If you implement AI on your existing motions, and your existing motions are wrong, all you do is speed up getting to the wrong answer. And you do it at a scale no single person working alone ever could.
This is exactly the blind spot the AI Revenue Risk Assessment is built to catch — not whether you’ve adopted AI, but whether what it’s doing right now is actually working: see how it works →
The Question Nobody’s Set Up to Answer
Put “prove the ROI” in front of most GTM leadership teams and the honest answer stalls almost immediately. Not because the AI isn’t working. Because nobody can say with any confidence whether it is, which is a different and more uncomfortable problem. Proving it requires knowing every tool that’s actually live, having a baseline to measure against, working from data your teams actually agree on, and someone whose job it is to own the number when it’s asked for. Most companies can’t check all of those boxes.
That gap is what the rest of this series is about: The specific, unglamorous reasons a well-adopted AI stack still can’t produce a straight answer to the ROI question, one at a time.
Speed without proof just gets you to the wrong answer faster.
This week, put one question in front of your GTM leadership team: name every AI tool live in your sales, marketing, and CS motions right now, and tell me what each one returned last quarter. If the room can only answer the first half, you already know where you stand, and it’s not as far ahead as the last two years of adoption numbers made it feel.
Post 2 in this series: Why most GTM leaders can’t name every AI tool already running in their own sales, marketing, and CS motions — and what that blind spot costs.
This opens an eight-post series on why most B2B SaaS companies can’t answer that question, and what closes the gap. If you want a scored, five-minute read on where your own organization stands, the AI Revenue Risk Assessment gives you the diagnostic: take the assessment →
Andrea Mulligan is a B2B SaaS executive and advisor with 30 years of experience building Customer Success, Professional Services, and GTM organizations. She works with PE-backed and growth-stage companies on CS transformation, revenue retention strategy, and post-sale model design. Start a conversation →