AI and automation are transforming agency operations, forcing agencies to restructure their teams and rethink their value proposition in an AI-driven world.

"The things that technology has made easier were the things that were actually quite easy," said Liane Nadeau, chief investment officer at Digitas, speaking with our analyst Nate Elliott at EMARKETER’s Ad Buyer Strategies Summit. "What technology has also done is just accelerated the pace, the demands, the needs of the business. And so it's become, yes, maybe a little bit easier in some senses, but so much more complex and so much more challenging."

Liane Nadeau speaking at EMARKETER's Ad Buyer Strategies Summit

As algorithms handle more tactical work, agency roles are evolving rather than disappearing. Nadeau compared the shift to factory automation: manual labor decreased, but new roles emerged to build, monitor, and quality-check the machines.

"The hands on labor of bids and optimization daily, and those budget moves maybe became more automated," she said. "But all of a sudden, you needed someone to build those machines. You needed someone to monitor and watch those machines."

This means staff focus must shift from tactical execution to strategic thinking. Agencies can no longer differentiate solely on optimization capabilities, platforms like Performance Max handle that better than humans.

As Elliott put it, “One of the concerns I hear is in a world where robots make a lot of these decisions, won’t everything just come out looking the same?”

Nadeau argued differentiation now comes from the strategic guidance and consumer experience planning that prevents AI from simply "creating more of the same."

What isn’t being automated yet

Even as AI takes over more executional work, some parts of media planning remain difficult to automate. One major obstacle is the way media budgets are still organized, broken into rigid, channel-specific line items that limit flexibility and slow down real-time optimization.

Nadeau pointed to two key barriers holding agencies back from fully fluid, AI-driven budget allocation:

• Infrastructure requirements: Agencies still sign contracts for inventory, which creates friction when reallocating budgets between channels.
• Organizational structures: Both agencies and clients require approval processes when moving significant amounts of money between traditional budget categories.

"We're getting to a place where we can do that a lot more rapidly," Nadeau said. "But we're not to a point, and I don't think we should be at a point where we just dump the money into one platform and say, spend it however you'd like."

AI as a channel requires dual-audience planning

“Obviously, more people are using AI every day, and it’s becoming a bigger part of the customer journey,” Elliott said.

This means agencies must now consider AI platforms as a channel where consumers discover and evaluate brands, creating a fundamental shift in how media plans are constructed.

Agencies are now building content for two distinct audiences: human consumers and the large language models (LLMs) that consumers increasingly use for research and decision-making. Brands must ensure they're discoverable when consumers ask AI tools for recommendations, or risk being excluded from consideration sets entirely.

"We're in phase one of LLM use by consumers, because I'm typing in ‘what are the best running sneakers for me’ and it's giving me some answers," Nadeau said. "I think we're going to very quickly get to a phase where I just say ‘buy me the right sneakers for me.’ And it goes off and makes those decisions."

Current advertising within chatbots like ChatGPT and Copilot remains in the learning phase, with measurement still too nascent to provide clear ROI. The more pressing need is ensuring organic discoverability through what's being called GEO (Generative Engine Optimization), creating content across paid, earned, shared, and owned channels that LLMs can crawl and reference.

Brands need to act now to establish their presence in AI-powered discovery, before consumers move from using AI for research to delegating purchase decisions entirely to AI agents. This requires thinking beyond websites to create a comprehensive content ecosystem that trains LLMs to recommend specific brands.

Watch the full session

We prepared this article with the assistance of generative AI tools and stand behind its accuracy, quality, and originality.

This was originally featured in the EMARKETER Daily newsletter. For more marketing insights, statistics, and trends, subscribe here.