AI Didn't Change Your GTM Motion. It Just Made It Faster.

PE-backed companies are all saying the same thing: "We've added AI to our GTM." Ask what changed, and most of the time the answer is a faster version of their legacy motion.

Faster isn't the same as different, and faster can't fix broken. Souping up the big block in a 1969 Chevy Chevelle means very little if the car is on blocks with no tires. This is the gap where most of the current AI-in-GTM spend is going to waste.

The Five Moments Worth Rebuilding, Not Speeding Up

SAP's Jan Gilg made this case clearly in a recent piece on SAP's News Center. Most GTM organizations don't have an AI problem, he argues; they have a decision-flow problem. He points to five specific moments in the customer journey where AI should change the actual mechanics, not just the speed:

Segmentation shifts from static firmographic data to real-time business signals. Engagement moves from higher-volume outreach to higher-relevance outreach.

Deal execution targets the invisible friction inside a company's own approval and quoting process, not the buyer's hesitation. Post-sale handoff gets an "account brain," a running repository of context that survives a rep leaving or an account transfer. And retention shifts from reactive account reviews to continuous, always-on scoring of expansion and churn risk.

None of that is a chatbot bolted onto an existing workflow. It's a completely different way of thinking about GTM.

The Bar: 40,000 Transactions a Year

Gilg's proof point is concrete enough to use as a benchmark. Amadeus, working with SAP, deployed an autonomous agent that reconciles unstructured payment data, clearing roughly 40,000 incorrect transactions a year that used to require manual fixes. That's the bar. Forty thousand things a year that no longer require a human to notice, escalate, and fix.

Why the Infrastructure Question Is Becoming a Diligence Question

AI in GTM isn't just an operating question anymore. It's a valuation question. PE firms used to price ERP chaos into a deal — messy back-office systems, no single source of truth, a lower offer to cover the cleanup. Clay's recent piece on AI-enabled GTM for private equity argues that the same discount is coming for GTM infrastructure. Buyers used to ask why the pipeline numbers were the way they were. Now they're asking whether the GTM infrastructure is AI ready or whether it needs to be built from scratch after close.

Both pieces hit the same point. Clean, enriched data beats a flashy agent every time. Autonomous outbound tools still aren't reliable enough to run without supervision. And the real advantage compounds across a portfolio, not inside one company.

The Same Pattern Is Hitting How Software Gets Built, Not Just How It's Sold

Zoom out further, and this isn't only a GTM-motion problem. It's showing up in how SaaS companies position their entire product against AI itself.

Kyle Poyar recently reviewed how 50 prominent SaaS and AI-native companies talk about competing with Claude directly. Thirty-four of the fifty have already shipped MCP connectors, making their own product accessible as a plugin inside Claude rather than competing with it head-on. Eighteen have published nothing addressing the comparison at all. Only twelve name Claude directly on their own site, and of those, just four go so far as to publish a real cost comparison between building in-house and buying their product.

That's the product-side version of the same question GTM leaders face operationally: is your value proposition something a general-purpose AI could plausibly absorb next quarter, or is it durable underneath that? The same tension is showing up on the content side. Anthropic's move to automatically watermark AI-generated text, prompted by EU transparency requirements, has already sparked pushback from users worried that lightly-edited AI-assisted work will carry a permanent, invisible flag.

The Actual Question

So the question isn't whether your GTM org has "added AI." Almost everyone can say yes to that now. The real question is whether your team is doing AI, or still doing the old motion, just faster.