The agency that changed its offer because of AI
I've had a wide range of AI conversations with agency owners this month - and the split has been striking.
On one side, people who are excited and testing Claude's new "Design" feature and 4.7 model. On the other, agency owners worrying about client data privacy, shadow AI and potential agency headcount reduction. The speed and nature of AI adoption varies widely too.
Three agencies, three different stages
In my new business calls, one medical communications agency owner told me "some of our clients don't allow us to use AI at all" - because their compliance teams have put the brakes on completely. Meanwhile, in my Account Accelerator cohorts over the last three months:A UK brand strategy agency has consciously decided not to adopt any AI in their creative development or account management workflowsA US full-service digital marketing agency in the travel sector has hired a dedicated "Head of AI Strategy" - and their client success team are using Google's NotebookLM for knowledge sharing and reducing adminA UK app development agency is building a networking group of CTO clients to increase their strategic value (because they can see deepening client relationships will become even more important)
One agency shares their experience
UpGrowth Digital shared their experience after 18 months of rebuilding workflows for AI. It's a 32-person agency serving 150+ clients and two years ago 70% of team time went on production work and 30% on strategy, client calls and relationship building. Today that ratio is 40% production, 60% non-production. Same headcount. Different work. Their weekly client report preparation time dropped from 3 hours to 25 minutes. Long-form blog turnaround from 8 working days to 3. They also spent six months hiring "AI-fluent" mid-level marketers before realising the trait that actually mattered was judgment - not prompting skills. Read the full story here →
Strategic ABM redesigned what they sell, not just how they deliver
Eloise Todd and I talked about this on the podcast recently. Her account based marketing agency didn't just use AI to speed up existing delivery. They built an AI-powered self-serve platform called ABM OS and launched a tiered model - full consultant delivery for clients who want it, self-serve for teams with the capability. The Account Director role at Strategic ABM changed as a result. Eloise is involved earlier in new business calls. She provides more consulting on which tier is right for which client. The strategic value of her client conversations has gone up.
Here's a summary of the main points we covered.
Or listen to the episode here →
The future of account management
The commercial point of having account managers has always been to make the agency's forecast more predictable - keeping clients longer and growing accounts through solving more of their problems. The whole point of AI handling the backend admin and production work is so account managers can finally do what they were always supposed to do: research their clients' business and market properly, prepare new ideas before client meetings and input into the monthly forecasting meeting with numbers that aren't a work of fiction. What would your top three growth accounts look like if you had three hours back in your week and had time to be more proactive?
Two questions to ask yourself
Are you adopting AI sporadically across different departments, or redesigning how work moves through the whole agency with an AI-forward approach?Are you listening to clients' changing needs, thinking about how the value you offer needs to change, evolving how you price and updating AI use policies as you go? (Alex, CEO of Strategic ABM calls it 'meeting clients where they are')
What to do about it this month
Here are some suggestions for actions you can take if any of this resonates:
If you lead an agency:
Ask your account managers to track a week honestly. You'll probably find 60-70% of their time still goes on tasks AI could assist with.
Pick one workflow and redesign it from scratch as AI-first. Client reporting is often the biggest win. Not "AI-assisted" reporting. A rebuilt report where AI automatically pulls data, structures it and then sends a draft to the account manager to review before it goes to the client.
Be explicit about where the saved time goes. Hours freed up go to client research and forecast-firming work - not more admin.
If you're in account management:
Work out which two or three tasks eat most of your week and ask your agency leadership to let you pilot redesigning one. Show what an hour saved looks like in commercial terms.
Use the time you get back for one thing: preparing a growth hypothesis for an account you lead that has the highest growth potential. Research the client, form a point of view about what you can offer next, share it with your leadership team for feedback.
Develop your judgment, not just your prompting skills. Knowing when AI output isn't good enough is the skill that matters - if you send an email that sounds like it was written by a lawyer rather than someone from a trendy performance marketing agency, it can change the whole dynamic of your relationship.
This is the hope for AI adoption
The purpose of account management hasn't changed. It's still to retain and grow existing clients and firm up the forecast. What's changed is that AI is starting to reduce the time spent on busy work.
