Over the last few weeks I've written three articles on AI. One on what it looks like inside a creative agency, one on where it can help with everyday admin and one on the mix of excitement and fear it brings. I'm also aware there's a lot of AI fatigue out there, so after this one I might back off the topic for a bit.
First though, there's a question I keep coming back to that I haven't covered.
Clients are asking agencies about AI more and more. Are you using it, where are you using it and what's your policy on it. What I haven't heard many of them ask yet is the follow-up. If it saves you time, when do my quotes come down?
I suspect that question is coming. It also made me ask something I don't think enough agencies have thought through. If AI saves time, where does that time go?
Three places the time can go
Saved time doesn't disappear. It ends up somewhere, and there are really three options.
- Back to the client - lower quotes and fees
- Into the agency - better margin
- Nowhere you can see - it gets absorbed into the job
Most agencies haven't chosen between them. That means the third one is the default.
Start with staying on budget
Before getting into pricing, I'd look at overspend. A lot of agencies lose margin because jobs overrun, not because the original price is always wrong. Some of the time behind an overrun is admin, like chasing approvals, writing up notes and pulling together status updates. That's the sort of work I covered in my second newsletter.
In my experience though, the bigger cause is usually creative over-delivery. One more route, one more round, a bit more polish because the team wants the work to be great.
If automation and AI take some of the admin off the team and you keep a closer eye on over-delivery, more jobs may land on budget. That protects margin without a difficult conversation with a client, so it's a good place to start.
Will clients ask for lower quotes?
Some will. Clients are having their budgets squeezed too and are being expected to do more with less, and naturally that gets passed on to their agencies. Add AI to the picture and I'd expect more of them to ask whether it's making things faster and cheaper. They may not wait for you to bring it up.
How hard that conversation is depends on how you price. If you charge by the hour and AI saves hours, your revenue goes down with them. That's simple maths, but it catches founders out. If you work on fixed fees or retainers, you're in a different position, because the client is buying an outcome rather than a number of hours.
Either way, it's better to know your answer before the question arrives.
Should agencies keep the profit?
I think they should, or at least a fair share of it. Agencies are businesses too. They have salaries, rent, software and everything else to pay for, and in my experience margins are often tighter than clients realise.
Learning to use AI well costs money. There are the tools themselves, the training and the time people spend working out how to use them, which often isn't chargeable to anyone. If every hour saved goes straight back to the client as a lower price, the agency never gets that investment back and has nothing left to fund the next improvement.
A healthy, profitable agency is better for clients too. It can keep good people, keep investing and keep getting better at what it does.
Clients rarely see timesheets or hours, so in most cases they won't know how long a job really took. What they do see is the fee. They remember what a similar job cost last time, and they may well ask what AI has changed. Even if they never see the hours, I'd think about trust. If a client finds out later that a job cost you far less than they assumed, how does that conversation go?
A middle route is to share some of the benefit with the client, keep some as margin and put some back into the agency. Whatever you choose, decide it before the client asks, not in the meeting.
Parkinson's law is the bigger challenge
Parkinson's law says work expands to fill the time available. Give someone a week to do a task and it takes a week. Give them two days and it often gets done in two days.
In a creative agency it usually shows up as over-delivery. A job is budgeted for a set number of hours and somehow that's how many get booked. One more route, one more round, a bit more polish. Creative work is never really finished, it just gets stopped.
Here's what that means for AI. Say a job is budgeted at 40 hours and AI saves a few of them on the first draft. If the team still books the full 40 hours, because the saved time goes on extra rounds and polish, the job costs the same as before. The agency has paid for the tools and the training and seen nothing back, so it looks like AI isn't making a difference.
It probably did save time. It just got spent.
How to keep the time you've saved
None of this is complicated, but it does need someone to own it.
- Agree what done looks like before the job starts - a clear brief and a clear definition of finished make it easier to stop when the work is good enough
- Re-estimate using what jobs now take - if your budgets are built on old actuals, they still contain the old time
- Check budget against progress part way through the job, not at the end - a ten minute catch-up while there's still time to change something
- Treat a budget as a limit rather than a target - and make sure finishing well, and finishing early, is recognised
- Keep timesheets accurate - if they're patchy, you can't see any saving
Not all of it has to be profit
I talk a lot about margin, but it isn't the only place for the time to go. Some of it can go into training, into bringing newcomers through, which I looked at in the third newsletter, into thinking time on a job, into time with clients or into your team not working late to hit a deadline.
Those are all good uses. They just need to be a choice, and not something that happens because nobody was watching.
So where will your time go?
If a client asked you tomorrow whether your quotes were coming down, what would you say? Here's what I'd want an answer to first.
- Where are we saving time, and do we know by how much?
- Do we charge for time or for outcomes?
- What will we say when a client asks about price?
- How much do we share, how much do we keep and how much do we reinvest?
- How will we stop the saved time being spent?
Where the value will be
There's a bigger point behind all of this. Clients are being squeezed and AI is making executional work faster and cheaper. In reality I think agencies will have to offer more strategic work, because a lot of the executional work will be lost to AI, outsourcing or both. If something can be produced quickly and cheaply, clients will expect to pay for it that way.
Strategic thinking is different. It comes from judgement, experience and understanding a client's business, and it's not something a tool or an outsourced production team can give them.
So as well as asking where the saved time goes, ask whether it's moving your team towards the work clients value most.
Clients should value your strategic thinking. That's where the value is, and it's where you get to set the price.
