What fifteen years of GTM technology taught us, and why the lesson matters more with agentic AI.
Every few years, go-to-market makes a serious technology bet. Marketing automation, then data platforms, then revenue intelligence suites. Each arrived on a clear and reasonable promise: more pipeline, a tighter forecast, stronger retention. Leaders invested accordingly, and they were right to. This year the bet is AI, and it is the biggest one yet, arriving with the highest expectations and the most stretched resources of any wave so far.
The useful question is not whether that spending was worth it. It is what all those waves taught us about the distance between a technology’s promise and the value it actually delivers, and whether we are about to apply that lesson to AI or spend straight past it.
The lesson was never the tools
In 2011 there were about 150 marketing technology products. By 2025 there were more than 15,000. And by Gartner’s measure, organizations use less than half of the capabilities they pay for. It is tempting to read that as waste, but that misses what happened. The technology usually worked. The promise was kept where the system underneath it was ready, and it stalled where that system was not.
The pattern in what stalled was consistent, and it was rarely the software. It was the readiness beneath it: data that had not been connected, definitions that had not been reconciled across teams, and people already stretched too thin to run one more system well. A capable tool on a shaky foundation does not fail loudly. It produces confident-looking output and quietly underdelivers on the promise that sold it.
AI changes what the lesson costs
Every technology before AI mostly reported. It scored a lead, drew a chart, flagged a risk, and then waited for a person to act. That pause was where a shaky assumption got caught before it did any harm.
AI does not pause. Given a goal, it decides and acts. The same unreconciled definition that once produced a misleading dashboard now produces an action, taken before anyone reviews it. The distance between a promise and the readiness underneath it used to be separated by a human and business context. Now it is separated by nothing.
The readiness that matters is agreement
Of everything the underlying system needs, the piece that matters most is the least visible: agreement on what the words mean. Take one word. *Engaged.* To marketing it means activity on its channels. To sales it means a real person returning calls. To customer success it means the product is being used deeply enough to renew. One word, three readings, one account. *Qualified, healthy, at risk, good revenue* each fracture the same way.
You can test this in your own business in an hour. Pick one active account. Separately, ask your marketing, sales, and customer success leads to describe it in a word and name the next best action. You will get three different answers. That is the readiness gap, and it is exactly what AI is about to act on, automatically, before anyone reconciles it.
Same budget, opposite outcomes
Follow two companies forward. Same budget, similar tools. One has that shared foundation in place. The other does not. Both rise at first, because AI delivers early efficiency that is easy to celebrate and easy to mistake for value. Then they part. On a shared foundation, AI compounds, because everything acts on the same picture. Without it, AI does not simply plateau. It accelerates the gap, and the early efficiency is what hides it. Same investment, opposite outcomes, decided by what sits underneath rather than by the AI.
The move before the next purchase
None of this is new. A shared definition across functions, clean and connected data, and people who can direct the work are requirements the last fifteen years already taught us to expect. What is new is that AI raises the cost of not having them, because the system now acts on whatever is underneath it.
So the first move is not another purchase. It is to make the words that drive revenue mean one thing across marketing, sales, and customer success before the AI is switched on. It is unglamorous, it is mostly not technology, and it is the readiness every promise has assumed and few have built.
We have already paid for this lesson, wave after wave. The only question left is whether we capitalize it before the AI starts acting, or after.