The Revenue Function Rebuilt: What AI changes about how businesses actually grow

Most businesses are trying to bolt AI onto a revenue organisation that was designed for a world without it.

I think that may be the bigger problem.

Not the tools. The organisation underneath them.

The old structure was built around scarcity

For most of my career, revenue organisations were built around a simple constraint: producing good information took time and people.

Someone had to pull the numbers. Someone had to reconcile them. Someone had to analyse them. Someone had to turn them into a deck.

Forecasting happened monthly or quarterly partly because that was how long it took to get a reasonably clean picture of what was going on.

We built teams, processes and layers around that reality.

AI is beginning to change the constraint itself.

Reports that took days can increasingly be produced in minutes. Account research can happen before a salesperson has opened their laptop. Forecasts can be updated continuously rather than reconstructed once a month.

And yet, in many organisations, the structure around all of this hasn't really changed.

Same organisation. Faster tools.

The most common response I see is to give people AI tools and ask them to become more productive.

The analyst still produces the report, only faster.

The planner still builds the deck, only faster.

The salesperson still prepares the proposal, only faster.

There is obviously value in that.

But I'm not sure that's transformation.

If AI materially reduces the time spent producing information, perhaps the more interesting question isn't how much faster people can do their existing jobs.

It's what those jobs should become.

From production to judgment

This is where I think revenue organisations start looking different.

Forecasting, for example, becomes less of a monthly event and more of a continuous signal.

The leadership conversation changes from:

“What happened last month?”

to:

“What's changing right now, and what are we going to do about it?”

That's not simply a better dashboard. It requires a different management rhythm.

Roles change too.

The analyst doesn't necessarily disappear. But perhaps less of the job is producing the report and more of it is understanding why something doesn't look right.

The planner spends less time assembling the standard deck and more time deciding which customer needs a non-standard approach.

The salesperson spends less time researching an account and more time understanding the politics, motivations and relationships inside it.

The common thread is judgment.

I wrote about this a couple of columns ago in the context of salespeople. The more I think about it, the more I believe the same shift applies to almost every function surrounding revenue.

And then there is headcount

This is where the conversation becomes more uncomfortable.

When AI removes manual work, the immediate instinct is often to ask:

How many people do we still need?

That's a legitimate business question.

But there is another one worth asking:

What could these people be doing that we previously never had enough capacity to do?

Deeper account planning.

Identifying renewal risk earlier.

Spending more time with customers.

Working harder on expansion.

Improving negotiations.

Understanding why a large opportunity isn't moving.

These aren't new problems. Revenue organisations have always needed to do them better.

We just spent enormous amounts of human capacity producing the information required to get to them.

That's why I think some of the most interesting AI conversations won't ultimately be about headcount reduction.

They'll be about headcount reallocation.

Not simply using AI to run the old organisation cheaper, but using the capacity it creates to build a better one.

Our incentives may be outdated too

There's another part of the revenue machine that hasn't received enough attention: compensation.

A lot of sales structures still reward activity.

Meetings booked. Opportunities created. Proposals sent. Pipeline generated.

But AI is making activity increasingly cheap to produce.

It can research hundreds of accounts, personalise outreach, draft proposals and generate follow-ups at a scale that would have been impossible a few years ago.

If volume is no longer scarce, should we continue rewarding volume in the same way?

Perhaps compensation needs to move further toward the things that still require harder judgment:

Retention.

Expansion.

Quality of revenue.

Customer outcomes.

Deal quality.

I'm not suggesting activity suddenly stops mattering. It doesn't.

But compensation is a signal. It tells an organisation what it really values.

And if the nature of valuable work is changing, eventually the incentives have to change with it.

This is the difficult part of AI transformation

Buying technology is relatively easy.

Redesigning an organisation isn't.

It means changing roles. Retraining people. Changing management rhythms. Rethinking incentives.

And sometimes having difficult conversations about which skills become more valuable as the work changes.

That's considerably harder than announcing an AI initiative.

Which may be why so much of the conversation still focuses on the tools.

But after spending much of my career building, scaling and restructuring revenue organisations, I'm increasingly convinced that the biggest impact of AI won't be visible in the technology stack.

It will be visible in how the organisation itself works.

Fewer hours producing information.

More time interpreting it.

Less time reviewing what happened.

More time deciding what to do next.

And incentives that increasingly reward outcomes rather than activity.

We've spent the last few years asking which AI tools revenue teams should adopt.

Maybe the more important question now is:

Does the revenue organisation we're putting those tools into still make sense?