I’ve been thinking about what Agentic AI means for consulting and agencies.

Every previous wave of automation has ultimately helped the industry. We automated execution, then moved further up the value chain. More strategy, more judgement, more advisory.

This time feels different.

Agentic AI isn’t just automating execution. It’s starting to automate judgement.

The pattern I’m seeing is that clients deploying service agents internally are getting very fast ROI from work that would previously have consumed agency hours. If that’s right, the implications for firms built on time-based billing and delivery volume are significant. Those economics become increasingly difficult to defend.

That makes me wonder whether simply bolting AI onto existing service lines misses the point.

The firms that come through this strongest may be the ones that rebuild around the things AI can’t easily replace - relationships, context, accountability and trust. The willingness to stand behind an outcome when something goes wrong. That’s fundamentally different from selling hours.

It also raises an interesting pricing question.

As AI pushes capability into commodity territory, where does margin actually end up? My instinct is that we see the same barbell effect emerging across professional services that we’ve seen elsewhere in the economy - margin concentrates at the lowest-cost, highly automated end, and at the highest-trust advisory end. Everything in the middle gets squeezed.

If that’s where we’re heading, a lot of firms are currently positioned in exactly the wrong place.

Curious where others are placing their bets.

#AgencyModel #Consulting #AIAgents #ProfessionalServices