AI without the headache: the rule of three for consulting firms in 2026

If you're running a consulting firm and you've spent the last 18 months running pilots, evaluating vendors, asking the team to 'experiment with AI'... and you're privately wondering why none of it has actually shipped... you are in extremely good company.

We closed ConCon26 with our Chief Product Officer Ben Jervis on exactly this topic.

Ben spent six years at one of the Big Four before joining CMap, and now leads our product team building CMap Intelligence - the AI layer turbocharging internal operations across 700+ consulting firms. Tom Rains, who has been with CMap for 13 years, walked through the product live.

The punchline of the session is the simplest GTM thread I've heard for AI in consulting all year. Here it is:

‍

1. AI fatigue is real - you're not imagining it

AI offers significant benefits. But for most consulting firms, the day-to-day reality is fatigue, confusion, fear, and a growing AI hypocrisy - pitching AI transformation to clients while running their own firm on spreadsheets.

"AI fatigue is real, and you're not imagining it."

• Stop running pilots that have no shipping deadline

• Audit which AI experiments in your firm have actually changed how you work

• Be honest about the cobblers'-children risk in front of your clients

‍

2. The rule of three (and why most AI experiments stall)

Ben grounded the framing in cognitive load theory. A University of Oregon study showed most humans only have the capacity to focus on three to four things.

AI doesn't escape that constraint, because humans still have to design, build, test and own every initiative.

"Cognitive load theory says human working memory has limited capacity. Once you go beyond three, the human brain really starts to struggle."

• Pick three AI initiatives. No more.

• Be a finisher. Get those three live and operational before adding the next three.

• Treat your AI shortlist like a board paper, not a roadmap brainstorm

‍

3. The impossible choice - internal ops or client delivery?

Most consulting firms are sitting with the same problem. Three internal-ops AI initiatives (resourcing, finance, ops) plus three client-delivery AI initiatives (new services, augmented offerings, vertical-specific tooling) is six.

Six is too many. So most firms quietly drop the unsexy ones - and then end up regretting it.

"Six initiatives. As soon as we go beyond three or four, we've got a problem. That's where AI headaches start all over again."

• Don't drop internal-ops AI in favor of client delivery - both compound

• Map your three internal-ops AI initiatives and three delivery AI initiatives explicitly

• Make the hard call on what stays and what waits - and ship the three that stay

‍

4. You shouldn't need to choose

The resolution? Adopt CMap for internal operations. Inherit the AI on resourcing, finance, ops, expenses, emails, cash flow. And free your three AI initiatives for the client-delivery work that compounds your firm's differentiation.

"Adopting CMap for internal operations means you can be free to focus on those exciting AI initiatives on the client delivery side."

• Identify which internal-ops processes are eating your team's bandwidth - resourcing, expenses, email filing, billing, forecasting

• Evaluate whether a single operating layer (with AI built in) can replace the patchwork

• Re-allocate the AI investment freed up into client-delivery initiatives

‍

5. The four non-negotiables for evaluating AI

If you're sitting on three vendor pitches this quarter, Ben's four non-negotiables are a useful filter.

Simple by design. Embedded in the platform you already run. Problem-led, not AI-for-AI's-sake. Beta-co-built with customers.

"We've been releasing all of our features in beta, so we can actually work in partnership with all of our clients."

• Test every AI pitch against the four non-negotiables

• Ask vendors how their AI is embedded in the platform you already use - versus added on top

• Walk away from any vendor pitching AI as a separate product SKU with a separate price tag

‍

Final thoughts

The single most useful thing about Ben's framing is that it gives founders explicit permission to stop trying to do everything at once.

Pick three AI initiatives. Make them client-delivery initiatives, because that's where your firm's differentiation lives. And let an operating layer like CMap do the AI work on internal ops underneath, so the three you pick are the three you actually ship.

That's what 'AI without the headache' means in practice - not less AI, but less choice, less waste, and more shipped.