What 12 CROs told us at our private “The Split” dinner

Matt Milligan
Co-Founder, Uhubs

We recently hosted 12 CROs and VPs of Sales for a private dinner in London. No slides, no sponsors. Just an honest conversation about AI in their revenue organisations.

We call these dinners “The Split” because that’s the pattern we keep seeing: a widening gap between teams figuring AI out and teams still circling the same pilot from nine months ago.

The room included SaaS, cybersecurity, and fintech leaders. Here’s what came up.

AI doesn’t fix prioritisation. It exposes it.

This was unanimous. The biggest AI challenge isn’t tools. It’s sequencing.

One leader described layering tools for months without fixing the data underneath. Faster dashboards, but full of numbers nobody trusted.

Without clean data and a clear operating model, AI just becomes an expensive magnifying glass for your existing problems.

The CEO matters more than the tech.

Every team making real progress had one thing in common: the CEO was personally driving it.

Not delegating to a task force. Personally in it.

One CEO blocks two hours a week for every employee to experiment with AI. No agenda, no deliverable.

That company had already replaced an expensive comp platform with a tool the reps built themselves. Because they built it, they actually use it.

Anyone who’s rolled out a tool the sales floor didn’t ask for knows how rare that is.

Build beats buy in more places than expected.

Multiple leaders have quietly replaced expensive point solutions with lightweight internal tools.

Comp calculators, deal prep assistants, proposal generators — things that used to cost six figures.

The ROI isn’t just cost. It’s adoption.

Reps use tools they feel ownership over. Vendor-imposed tools become shelfware within weeks.

The competitive advantage isn’t which tools you buy anymore. It’s how fast your team can iterate on what they’ve built.

AI increases capacity. It can also hollow out judgement.

This was the most debated topic.

On one side: AI clearly makes reps faster.

On the other: one CRO found reps sending proposals with incorrect positioning because they never reviewed the AI output — they just hit send.

Another leader from a large cybersecurity business argued that reps lose something fundamental when they stop manually working their own deals: the instinct for whether a deal is genuinely progressing or just moving through CRM stages.

The room didn’t solve this one. But the consensus was clear:

AI should augment preparation, not replace judgement.

Execution gaps matter more than idea gaps.

For global leaders, the problem isn’t knowing what to do with AI. It’s getting consistent execution across regions.

HQ rolls something out, adoption is patchy, and AI amplifies whatever operating model you already have.

There’s also a structural issue that doesn’t get enough airtime: most AI tools are heavily optimised for English.

Call analysis, coaching insights, and sentiment tracking all degrade outside English-speaking markets. AI-driven coaching is essentially unavailable for a huge portion of the global sales floor.

Efficiency is now a board-level mandate.

Several leaders were under explicit instructions: more revenue, same headcount or fewer.

One leader with an IPO approaching described 2× revenue without growing the team as his entire focus.

“Exploring” and “experimenting” aren’t acceptable answers to PE boards anymore.

They want revenue per head to move.

Teams that started early — even clumsily — can now point to results. The ones who waited for the perfect solution are scrambling.

The unsolved problem: what separates top performers?

Multiple CROs admitted they still spend huge amounts of time manually working with RevOps to understand why their best reps win more.

They can see pipeline data and activity metrics. But they can’t explain what behaviours separate the top 20% from everyone else.

This tracks with what we see at Uhubs every day.

The real gaps between top and bottom performers are rarely product knowledge or demo skills. They’re prospecting capability, commercial strategy, and market acumen.

The appetite isn’t for more AI.

It’s for clarity.

The adoption problem nobody’s cracked.

How do you get salespeople to use AI tools when they don’t want to?

High-tenure reps often see AI as surveillance, a threat to autonomy, or just unnecessary.

Several leaders have quietly given up on their most resistant people.

The problem is that selective adoption creates data gaps. Half the team on the tool, half off it. Outputs get unreliable.

Holdouts then point to unreliable outputs as proof they were right not to bother.

The leaders making progress flipped the messaging.

Instead of saying:

“This helps us track performance.”

They reframed it as:

“This helps you earn more with less admin.”

Simple reframe. Most rollouts still haven’t made it.

What we took away.

The winners aren’t the teams doing the most AI.

They’re the ones using it to sharpen judgement, not replace it.

Sequencing deliberately, building where it makes sense, and staying focused on what actually drives revenue:

• Clean data
• Strong coaching
• Understanding what makes your best people different

If you’re a CRO dealing with some version of these challenges, we’d like to have you at the next dinner.

Small on purpose. Honest by design.