AdAge Feature: How Agency Buyers Evaluate AI Claims in M&A Deals

Article was originally published in AdAge. Click to read the original AdAge article.

AI is everywhere—on earnings calls, in pitch presentations and every agency capabilities deck. Scratch beneath the surface and a different picture emerges. Recent research from the National Bureau of Economic Research suggests that CEOs, the people demanding efficiency gains from AI adoption, are among the technology’s least frequent users.

Those same CEOs are the ones representing their agencies during the M&A process. Over the past 12–18 months, the first question in nearly every agency sale process we’ve run has been: What are you doing with AI—and how does it impact future growth?

While every agency has an AI story, very few have evidence showing impact.

As an investment bank in the marketing and communications space, we review hundreds of agencies every year, and we’ve yet to see AI meaningfully change a valuation. It’s not because buyers don’t care. The absence of demonstrable economic value or impact means in most cases that AI is just window dressing, or worse, an over-promised cure-all for any client ailment.

AI use in day-to-day operations is table stakes today. While that alone certainly helps in conversations with potential acquirers, it isn’t a valuation driver.

Yet.

There’s no shortage of agencies positioning themselves as AI-enabled. New products. Proprietary tools. Rebranded methodologies. But buyers aren’t underwriting positioning. They’re underwriting cash flow. And when they look under the hood, two things tend to be missing: Talent and financial evidence.

Without either, AI is just a veneer. Which in itself raises a simple question: How do you separate real capability from well-packaged narrative?

Start With the Scientists

There’s a simple way to cut through the noise: follow the people. If AI is truly embedded in an agency, it should show up in the talent base. Not just strategists and prompt engineers, but people who actually build and operationalize systems. Machine learning specialists, data engineers, technical architects and security and compliance leads.

One CEO we work with has a straightforward rule when evaluating AI vendors: Ask who built the technology, and what they’ve built before. If the answer is vague, the conversation is over.

It’s a useful filter. Because without that layer of technical expertise, most “AI capabilities” are simply wrappers on top of third-party tools. Useful, in many cases, but rarely defensible and not a source of long-term value. Sure, that bright (and well-meaning) young strategist might have built an awesome vibe-coded landing page for your biggest client. But will that new page meet enterprise-level security and code requirements?

While AI continues to get smarter by the day, it remains but a tool. And a fool with a tool is still a fool, no matter how smart the tool is. All digital transformation requires people, process and technology. But in the case of AI, ironically, starting with the people has never mattered more.

Follow the Money

The second test is even simpler: where does AI show up in the numbers? If AI is driving efficiency, it should be visible in margins. If it’s creating new capability, it should be visible in revenue. In practice, we’re seeing very little of either.

Many agencies talk about productivity gains. Fewer can point to sustained margin expansion. If output per employee is increasing but margins remain flat, the value simply hasn’t been captured. Gross and net margins, utilization rates, revenue to comp ratios should all be moving in the right direction if AI is making a meaningful contribution.

On the revenue side, the story is similar. AI is frequently positioned as a new service layer, but rarely broken out in a way that shows attributable growth. There are few examples of standalone AI-driven revenue streams, new pricing models or clearly defined product lines that buyers can underwrite.

That doesn’t mean AI isn’t being used. It clearly is. But usage alone doesn’t create enterprise value.

Why Buyers Remain Cautious

This is why AI, despite the hype, is not yet a multiple-expander in agency M&A. Buyers want to believe in it. They are actively looking for platforms where AI can accelerate growth or unlock efficiency. But in the absence of evidence, they default to treating AI as a hygiene factor—a necessary capability, but not a driver of upside. In some cases, over-claiming can even have the opposite effect. When AI narratives outpace the underlying reality, it introduces risk. And risk, in M&A, means the selling price gets discounted.

One private equity principal I spoke to recently has gone no-bid on his last two marketing services deals, simply because of the AI question. Not seeing evidence makes it difficult to build, let alone support, a thesis.

What Actually Moves Valuation

None of this suggests AI won’t reshape the agency model. For agency owners looking to translate AI into enterprise value, the bar is higher than a capabilities slide or a new product name. If a serious investment has been made to advance AI in the business, show the people who are building and owning the technology. If AI is creating enterprise value then it needs to be supported by evidence of margin expansion, revenue growth or new commercial models.

AI may not be lifting valuations today, but it is now part of the diligence process and something agency acquirers expect to see, understand and pressure-test. Over the next few years the agencies that really do leverage AI will reap the rewards while the ones that say they do will reap only what they’ve sown. Provided, of course, that the claimed AI leverage is backed by evidence of real impact, in talent and in financial performance.

Until then, AI isn’t the key to a premium agency valuation. It’s the fastest way to separate real capability from very expensive snake oil.