Most AI services deals don't die in negotiation. They drift. The buyer stays interested but never commits. You keep showing up. Months pass. The deal never closes.

You’ve done everything fractional CROs preach: streamlined your CRM, clearly defined your opportunity stages, and created progression checklists to keep everyone accountable. The deals still stall between the steps.

Challenger Sale creator Matthew Dixon puts the share of complex B2B deals lost to "no decision" at somewhere between 40 and 60 percent. Not lost to a competitor. Not lost on price. Lost because the buyer couldn’t get to yes..

If you're running an AI enablement practice and you're honest about your pipeline, these are the deals that hurt most. The champion was engaged. The problem was real and there was a fit. And then, slowly, nothing.  The process became too drawn out, too demanding, too inconclusive. The champion ran out of internal energy before the deal reached a decision point. What killed those deals wasn't competition or price or fit. It was decision fatigue, and a process that never forced a conclusion.

Three things make this category different.

First, ownership of initiatives is genuinely unclear. AI enablement projects frequently end up with the CEO even in organizations with 500 or more employees. Not because the CEO is the natural buyer, but because no one else has authority. Budget lines rarely exist at the start of the year, and creating one mid-cycle is political work. And buyers think they understand AI better than they do, arriving with opinions that need to be revisited before real discovery can happen.

None of this is new if you've been reading this newsletter. But the remedy is underexplored.

So what’s the remedy? 

A successful software sales professional I was lucky to work with addresses the drift problem with a single move, three-quarters into the discovery call. When he feels the prospect is qualified, he pauses and says:

"In my experience, nobody buys technology without three things being true: first, there's got to be a business case for it; second, you need to validate that the tech works; and last, you need to make sure we're the type of company you'd like to do business with."

Then he proposes launching three workstreams – in parallel – to work through each one.

The logic holds for software. It holds even better for AI services, where the deal complexity is higher and informal drift is worse.

Here is how I would phrase it for selling AI services:

"Nobody commits to a project without ensuring three things are true: that there is a business case for it, that there is certainty around delivery, and that we're the type of company you'd like to do business with."

Three workstreams running in parallel. Same structure. Substantially different execution.

This workstream isn’t fundamentally different in services. You still need to understand your client’s business and show how your engagement moves the metrics they care about. Frame your impact in terms of revenue protected, cost removed, risk reduced, time recovered.

In practice, this means translating “we’re struggling with Copilot adoption” into “a more productive workforce enables 20% YoY revenue growth without increasing headcount.”

If your champion is a functional leader, the business case must withstand their CFO’s scrutiny. If your champion is the CEO, the path to approval may be shorter, but the numbers still need to hold up in the boardroom.

And if your champion is withholding information or refusing introductions to other stakeholders, the deal was never real.

In software, technical validation is often the “easy” step. A detailed demo or a short proof of concept brings technical leadership on board. In services, you don’t get to hand the buyer a sandbox.

What you can do is remove every uncertainty about whether your team will actually deliver. That means case studies from comparable projects. It means credentials and bios for the specific people who would work on this engagement, not a generic team slide that gets replaced by a different crew in week one. It means a methodology the buyer can examine: a structured system, not a collection of smart people improvising.

Buyers in this workstream are asking three things: Will you deliver what you're promising? Will the people we meet be the people who do the work? Can we actually work together?

Address all three directly. Don't wait to be asked.

This is vendor assessment in a softer wrapper. It’s a rite of passage for sellers to be burned by bringing a “done deal” to the pipeline meeting, only to be handed in a 257 question vendor onboarding questionnaire the next day. 

While ISO 27001 and SOC 2 are not yet universal requirements in services, that doesn’t mean you won’t be asked to prove you’re a safe partner. You still need to demonstrate that your business is financially stable, properly insured, and contractually sound.

The review will include credibility checks: reference calls with comparable clients, exposure to your broader team, and opportunities for stakeholders to assess your culture, communication style, and professionalism. These moments allow the buying group to de-risk the engagement, and they matter more than most founders realise.

This workstream is often skipped or rushed. That’s exactly why deals stall at final approval: the champion’s internal sponsor can’t confidently advocate for a firm the broader decision-making group has never properly evaluated.

There’s no doubt the script I’ve outlined is uncomfortable to deliver. If you’re a technical founder not used to selling — or British and instinctively allergic to confrontation — “uncomfortable” may in fact mean “terrifying.” 

That's the point. When you address the three workstreams head-on, you may discover the buyer isn’t ready to commit. That’s not bad news. It’s clarity. You’ve just disqualified a deal that was quietly occupying space in your pipeline and diary, keeping you away from other prospects.

When a buyer says yes to all three, something shifts. They've shown commitment. They're thinking about the purchase holistically, covering every base their organization will raise before signing. And because the workstreams run simultaneously, you're compressing a cycle that typically stretches across three to six months.

Most vendors wait for the buyer to steer. This framework puts you in front, handing the buyer a map, and asking them to start walking.

The champion who says yes to all three is the champion who can actually close.

— Daria