AI Won’t Kill the Agency Business. It Will Expose the Weakest Business Models.

We’ve heard the same warning so many times now: AI is coming for the agency business.

Every few weeks comes another prediction about how AI will replace creatives, vacate the account management department and automate strategy, turning the traditional agency model inside out.

But AI will not, in fact, kill the agency business. What it will do is expose which business models are most vulnerable.

And for investors, that distinction matters.

The agencies most at risk are not those with the fewest AI tools but, rather, those built around the work that AI increasingly can replicate.

On the other hand, the agencies that stand to become the most valuable are those that own something harder to reproduce—things like trusted client relationships, specialized expertise, proprietary data, diversified revenue, tech-enabled workflows and the ability to prove how advertising connects to business outcomes.

If I were an investor evaluating an agency today, these are the questions I would ask:

Who benefits from the productivity savings realized by AI?

Agencies have long operated under the same drill: A client comes in with a problem. The agency assigns people to solve it. More work equals more people, which leads to more hours and more revenue.

AI has thrown a wrench into that arrangement. Obviously, if a team can produce twice as much work in half the time, the old model becomes extinct.

But who exactly gets the benefit of all that improved productivity? Does the client get a price cut? Does the agency up its margins? Or does it reinvest the savings, and how?

Investors should demand a clear answer.

An agency that becomes more productive but cannot translate that into some real benefit is not a better business—it just does the same work faster.

That’s efficiency, but it’s not necessarily value creation.

Is the agency selling expertise or execution?

AI is tremendous at making production easier, doing everything from writing copy and creating images to accelerating workflow. But it also makes production less scarce, and when something becomes less scarce, its economic value typically depreciates.

In the case of expertise, the opposite is true. Deep knowledge of a category or a sophisticated understanding of a client’s customers is harder, if not impossible, to automate, making it infinitely more valuable.

The agencies that possess the strongest expertise will move ahead of the pack, as they spend less time selling sweat and more time selling judgment.

Does the agency own its AI-related IP?

When it comes to an agency’s AI investment, it is essential to ask one simple question: What technology does the agency actually own?

Simply adding an app that’s available to everybody else does not make a competitive advantage. Real value comes from what the agency builds around it.

Does the agency have proprietary models? Unique datasets? Tech that demonstrably improves campaign performance? Proprietary AI tools that make the client stickier?

If the answer is no, then an agency may be a user of AI but it’s not a master of it. There’s a huge difference.

Is AI really improving margins?

A lot of companies can point to AI-enabled productivity gains, but investors need to take a closer look at the economics.

Is headcount rising even as AI is making things more productive? If so, where and why?

If delivery costs are declining, where is the benefit?

If the company is spending aggressively on AI, when can investors expect to see a return?

The ultimate test is simple: Does AI make the company financially better off?

How diversified is the agency’s revenue?

It stands to reason that a vulnerable agency—say, one that is financially dependent on one or two large clients or one category—will be made even more wobbly by the ascent of AI.

An agency with multiple revenue streams—creative, strategy, media, consulting, commerce—will have more opportunities to use AI to create demand rather than just cut costs.

While diversification does not eliminate risk, it does make for a stronger agency.

 

So, again, the agency business is not disappearing, despite what the headlines suggest.

But it is changing.

AI will not kill the business, just as the shift to digital and programmatic advertising didn’t. AI will ultimately expose which businesses are made stronger and weaker because of the technology.

It’s not a question of how much work AI can be assigned to do. Rather, it’s how much value can be created as AI does more of the work.

That’s a very different equation.

The winners will not necessarily be those agencies with the shiniest AI tools but those that masterfully combine the latest technology with things that cannot be easily replicated by a machine: relationships, reputation and resiliency.

• Who benefits from the productivity savings realized by AI?
• Is the agency selling expertise or execution?
• Does the agency own its AI-related IP?
• Is AI really improving margins?
• How diversified is the agency’s revenue?

By asking those questions, investors stand to learn more about an agency than they will from the company’s latest press release about an AI tool it has acquired.

The real threat isn’t that AI will replace agencies—it’s that AI will make it impossible for vulnerable agency models to hide.

And that is very good news for the future of the industry.