Ask an agency leader how AI is changing their business and you’ll usually get one of two rehearsed answers: breathless optimism or quiet denial. Andreas Schwarz, Co-CEO at the Digital Experience Group, gives you neither.
“AI gets me out of bed in the morning, in a good and in a bad way. On the one hand, it’s definitely a massive change and a challenge for our business model. But it’s also an amazing opportunity.”
Schwarz leads a group that spans both ends of the digital spectrum: Ergosign, a company focused on user experience, and Small, a classic digital marketing and digital products agency that helps clients put their products in the best light through digital storytelling. That vantage point, with UX craft on one side and marketing production on the other, makes him unusually well placed to say what AI is actually doing to agency work, as opposed to what the hype says it should be doing.
What follows is his honest accounting: the real efficiency number (it’s not 10x), the outsourcing work that quietly disappeared, and why the most AI-mature clients are doing something unexpected: putting money back into brand. His perspective lands alongside other voices in this series, from Claneo’s Matthäus Michalik on why GEO is not just SEO to Slopelift’s shift from SEO to AI search.
The real number is 20 to 30%, not 10x
The most refreshing thing Schwarz says is also the least dramatic. Asked what efficiency gains AI actually delivers across the services his companies provide, he doesn’t reach for a moonshot figure.
“It really depends on the details, and you always have to double-click into stuff. You see areas where a task takes 20 to 30% of the time it took before. In other areas, it doesn’t have an effect at all. Overall, in the services we provide, it’s about 20 to 30%, more or less the average KPI you can read everywhere. Realistically, that’s also what we see.”
Two things stand out in that answer. First, the distribution is lumpy: some tasks collapse to a fraction of their former effort, while others are untouched. Second, the headline industry number, the 20 to 30% you see in every consultancy report, matches what a hands-on agency group actually measures. No inflation, no deflation. And Schwarz notes his teams tend to work on the high end of product work, where the easy wins are fewer.
Where the gains land is telling. “With a relatively small team, we can do a lot more. Replicating content, building content: we spend less time on the tedious, repeating jobs and focus more on the exciting part, where we can really make a change and our creative teams can play their strong hand.”
The work that used to go nearshore now stays in-house
Buried in the efficiency discussion is a structural shift that should worry outsourcing providers more than agencies.
“The tasks we would have outsourced into nearshoring, we can now do ourselves with AI. I’m not saying it’s black and white, but there are a couple of additional shades of gray in the game.”
This is the quiet reallocation happening inside agencies right now: the production layer that used to be arbitraged on labor cost is being arbitraged on compute instead. The work doesn’t leave the building anymore. For clients, the delivery looks the same. For the agency’s cost structure, and for the nearshore firms that used to catch that work, it’s a very different picture.
AI-mature companies are reinvesting in brand
Where are clients putting their budgets? Schwarz’s answer starts predictably and ends with a twist.
“I would say the majority of clients invest more in products than in brands,” he says. But that’s not the whole story. “There are a couple of them thinking about revamping the brand, and that’s where you see investment. It’s repositioning, rethinking.”
The twist is who those companies are.
“Especially companies who have taken the AI change seriously and are a little bit ahead of the curve, there you see reinvestment in brand again. There’s a clear picture that, yes, you need to differentiate.”
Read that back slowly, because it cuts against the intuition that AI-forward companies are the most product-and-performance obsessed. The companies furthest along the AI curve are the ones concluding that when production is cheap and content is abundant, brand becomes the scarce asset. When everyone can build and publish at machine speed, differentiation is the thing machines can’t commoditize. It’s the same conclusion Zurich Insurance Germany reached from the demand side: brand mentions are the new standard in AI search.
What separates those clients from the rest isn’t industry or size, in Schwarz’s telling. It’s maturity. “They take the data-driven, programmatic, AI-driven path a lot more serious, from the get-go, or have started earlier. It’s the AI or digital maturity they have.” Plus, he adds, the individuals: on every client side there are people who are “more visionary and more adventurous” in these areas, and they pull their organizations forward.
The Accessibility Act woke people up
Ergosign has always positioned itself as human-centric, so accessibility is not a bolt-on for Schwarz’s teams.
“Accessibility is playing a really big role in how we present content. Ergosign has always been human-centric. The human part has always been at the forefront of our thinking. The word itself and the offering didn’t exist ten years ago, but the ideas have always been there.”
What changed is the client side, and it took regulation to do it.
“From the client side, there’s more focus now. And with the push through the European Accessibility Act, that has really woken up quite a few people.”
It’s a familiar pattern: values-driven agencies advocate for something for years, and a compliance deadline finally converts it into budget. The agencies that treated accessibility as core practice rather than a checkbox are the ones positioned to serve that demand now.
“The relevance will be lower in the future, but I don’t see websites going away completely. As a basic platform to look for products, especially with strong products and strong brands.”
Note the qualifier, because it connects back to his brand argument: it’s strong brands whose websites keep their pull. The weaker your brand, the more your website’s fate depends on whatever an AI engine decides to cite. The stronger your brand, the more people come looking for you deliberately. Which is one more reason the AI-mature companies are reinvesting in exactly that.
The hardest problem isn’t the technology
Asked what the biggest challenge is right now, Schwarz doesn’t name a tool, a model, or a skills gap. He names the business model.
“The majority is really how to manage the change to our business model. There are opportunities and there are challenges: how to balance that, make sure we stay relevant in the market, and also find a proper commercial model which creates a win-win situation for us and our customers.”
That last part is the unglamorous heart of the agency AI story. When a task that took a day takes an hour, who captures the difference, the agency’s margin or the client’s invoice? Every agency is renegotiating that answer right now, client by client. The technology is the easy part. Pricing it is the transformation.
Key takeaways
Plan around 20 to 30%, not 10x. Real-world AI efficiency gains across agency services average 20 to 30%, matching the industry benchmark. The gains are lumpy: some tasks collapse to a fraction of their former time, others don’t move at all.
AI is eating nearshoring before it eats agencies. Production work that used to be outsourced on labor cost is moving back in-house on compute cost.
Watch what AI-mature companies do with brand. The clients furthest ahead on AI are reinvesting in brand, because when content is abundant, differentiation is the scarce asset.
Digital maturity, not industry, predicts AI adoption. The clients taking the data-driven, programmatic path seriously are the ones with existing digital maturity, plus a visionary individual pushing internally.
Regulation converts values into budgets. Accessibility was always the right thing to do; the European Accessibility Act made it a line item.
Websites will matter less, not disappear, and strong brands are the ones whose websites keep their relevance.
The real transformation is commercial, not technical. The hardest question isn’t how to use AI. It’s how to price work in a way that stays a win-win for agency and client.
Want to see where your brand shows up in AI answers, and who’s getting cited instead of you?Track your AI search visibility with OtterlyAI across ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, Gemini and Copilot.
