June 2026 · Anthony Sogg

You Use a Third of Your Martech. Adding AI Will Not Fix That.

The 2026 plan is to bolt AI onto tools you never operationalized. It is the same mistake as last decade, just automated and more expensive.

A while back I wrote that the enterprise martech stack had become a graveyard of good intentions, and the funeral keeps showing up with new vendors. I want to come back to that line, because the 2026 data turned it from a complaint into a warning.

Gartner's numbers: marketers now use about 33% of their martech stack's capability, down from 42% in 2022 and 58% in 2020. That is not a plateau. It is a slide. Roughly half the tools sit underused, and only about 15% of organizations qualify as high performers who actually hit their goals with what they bought.

The crows are coming home to roost. Many of these platforms and tools are the result of chasing after a new technology, a name, a buzzword, or the evidence of a very good salesperson. Those are not reasons for technology.

A clear business problem, a strategy to solve said problem, a goal to chase and stick with, a unified business effort. Those are reasons for technology trade ups or changes in business.

And yet we haven't learned. At. All.

The 2026 Gartner CMO Spend Survey has marketing leaders putting 15.3% of budget into AI while only 30% say they are ready to scale it.

So, the plan, said plainly, is to spend more on a new capability layer while using a third of the last one. We are not learning. We are accelerating.

In any other time period that would be grounds for dismissal.

AI does not fix the org chart

AI is not a fix for an underused stack. It is a multiplier on whatever is already there, the dysfunction included.

I have watched, advised, and cautioned enough CDP and orchestration implementations to see the same mistakes made repeatedly from miles away, and it is rarely the fault of technology.

The data flows in, usually just as messy and sometimes even messier than before. The segments build and problems compound. Nothing changes downstream, because the teams consuming the output never agreed on what they were trying to do. Suddenly, finger pointing becomes the new hot trend.

Drop an agent into that and you have not settled the disagreement. You have automated it, at machine speed, with less human attention on the result.

A tool you use at 33% does not have a technology problem. It has an ownership, data, and strategy problem. Nobody owns the marketing architecture, so capability accretes and utilization decays. The data's a mess, so the sold in capability is never achievable because the engine is being fueled with crude oil instead of race fuel. The strategy changes every quarter because leadership and the board are reactionary and can't see things through.

AI bought into that same vacuum will accrete and decay the same way, only faster and with a bigger invoice.

The unglamorous prerequisite

What fixes utilization is not a better tool. It is a single question asked of every tool and every agent before it gets a dollar: what customer question or business problem does this help us answer or solve that we cannot answer or solve today?

If the answer is "it is for reporting," that is a symptom, not a strategy. If the answer is "the demo was incredible," that is the fraud detector going off inside your own building.

Addition by subtraction, Occam's Razor, "simplify and add lightness". Pick your mantra; I'm partial to the latter as a Colin Chapman fan.

Operationalize what you have or retire it. Get utilization off the floor on the stack you already own, prove a team can run a capability end to end, and only then hand that team an agent to amplify it. And if they can't run capability end to end train them. Invest in their growth and capability before you stick an expensive flavor-of-the-week AI pet project on top of everything.

Amplify competence and you compound value. Amplify a vacuum and you compound the bill.

What 2026 will actually sort

The 15% of high performers are about to pull away, and not because they spent more on AI. Because they had an operating model AI could plug into. The other 85% are about to learn an expensive lesson: a capability you cannot operate is not an asset, it is a liability with a renewal date. AI does not change that math. It just makes the number bigger in both directions.

Buy the AI. I am not waving off the future. Just do not buy it to rescue a stack you never operationalized or a strategy you never followed, let alone developed. AI is good and getting better but it won't plug the gaps that you've left untended.

That part was always leadership and the organization's job. It still is.

Anthony

← More writing · If this was useful, or wrong in an interesting way, tell me: hireme@crmjedi.com.