Inside the messy middle of January Digital agency’s AI adoption

Brands like Coca-Cola have used AI to create constantly updated, hyper-personalized digital personas to scale creative from talent partnerships. Brands and their agency partners have started using agents to plan, execute, and optimize ad purchases.

At the same time, clients have high expectations, looking to AI to automate so much that it could start to cut into agency margins. Companies shelled out on AI tools while return on investments are still pending. To put some numbers to it, 40% of companies that measure their AI cost savings realized reductions of 10% or less — expectations were higher for more meaningful improvement, according to a new Bain & Co. global survey of large companies and reported by Bloomberg. 

January Digital finds itself in the messy middle, and Drabicky is quick to admit that. For the past two years, the agency has been navigating client expectations, the breakneck pace at which AI tools are moving and internal team capabilities to strike a balance that works for the agency. 

So far, it’s a combination of transparency with clients to manage expectations and convincing internal teams that they’re not building tech that’s going to put them out of a job, the agency founder said. 

“It’s working for now,” he said. “Ask me again in six months.”

This interview has been lightly edited for clarity. 

When AI automates the daily ‘grunt work,’ what does the actual work for junior talent look like, pen to paper?

We actually took the approach of giving our entire company Claude license and said, “You guys go build whatever you build. Let’s see what you build on top of the tools that we already have.” Obviously, there’s constraints and security and all that, but we wanted to see what people would naturally build to solve and make things better. 

One person did something very basic, where they automated how their pacing was built into decks and reported back to their client already. Someone else built this tool that takes all creative insights, compares them to the upcoming creative assets that are about to come out, gives it a ranking score of what it thinks will be there, and then automates all the reporting past it.

Basically each month we take the top two that are most universal. We processize them, put them into our tool set, let people go out and build more. We try to keep doing that until we run out of ideas. 


We see a lot of ‘anti-AI’ sentiment among Gen Z on social media, yet they are expected to use these tools in the office. How has your younger staff responded to that?

This is something we struggled with a ton. What we’ve been consistent on is, “If you continue to help us build great tools, what that means, is that you guys will continue to be here, we’ll be more efficient and will look differently, but we are going to continue to use you to help make sure that our tools are as best as possible.” Our team has really bought in on that. On our Mikey Intelligence platform [the agency’s proprietary, client-specific analytics LLM], we have 100% usage. On our [Mikey] OS, we’re like 92% usage. 

Is there anyone who’s declined or refused to use company AI tools?

Honestly, the bigger challenge is every client organization has a completely different openness to it and understanding of it. That’s where our jobs get harder. One client wants to be 100% AI, and the other one, not so much. That can cause some choppy waters that we’ve got to navigate with our teams. Otherwise, you get one team who’s super proficient and one that’s not, but you don’t want to just move people around. So we have to navigate that a little bit, but nothing too extreme yet. 

Has the financial model for testing AI evolved? Do clients have to foot the bill upfront, or are agencies absorb, baking those R&D costs into their broader fees?

I will tell you with our existing clients, and our Mikey Intelligence tool and all that, we rolled it out to all of them for free. We very openly said, “We’re unsure how much you guys will use this or what it will cost, so it is free. At some point, if it comes super expensive, we’re going to come back to you guys, but it is free for now.” For future clients, we look at them and go, “Look, here are the tools, here’s why you want them… If you don’t want to pay for this technology, the inverse is to go pay for more people, and more people is more expensive than this.”
So far, people go, “OK, I got it. I understand what you’re after.” Not to say that we haven’t had people say, “I don’t want to pay for it or any of those things,” that’s definitely happened too. But we’re still trying to figure it out and figure out the right balance. 


Looking at the big picture, how would you summarize the impact of the AI transformation on agency economics?

AI has added some efficiency, but those two things aren’t aligned yet. You can get into situations where clients are pushing you down for lower prices. If you say yes, you’re going to hurt your margin and you’re in trouble — or you say yes, you hope your AI tools catch up in the long term, and that’s the risky spot to be in. There’s a second current that’s coming in now where it’s like, “Oh, AI automates everything. I’ll just put it on this tool and let it run.” In some cases, you’re getting clients go, “I can pay you a million dollars, or I could pay this other tool $100,000, and if it’s $ 70% is good, fine.”


It is a little bit risky right now, because it’s very unclear on what reality is. The only approach we need to take is just be super upfront with folks on it — here’s where we make margin. 
Here’s where we don’t. Here’s where it’s adding efficiency. Here’s where it isn’t. These are the tools we use that we find to be really effective.