AI is changing how brands reach customers, with marketing leaders increasingly focused on how AI-powered search engines and agents discover, assess and represent their businesses, new research from Atlantic Insights and Contentful suggests.
The report surveyed 350 marketing decision-makers across the US, UK and Australia. All respondents worked for organisations already prioritising AI, meaning the findings reflect businesses that have already committed to the technology rather than the wider marketing sector.
The research identifies five areas that marketing leaders need to address: autonomy, judgement, legibility, measurement and ownership.
More than half of respondents (51%) said AI-powered search-and-answer engines and AI agents were equally important areas of focus. Twenty-nine per cent prioritised search-and-answer engines, while 19% put greater emphasis on agents.
Agentic commerce is also gaining traction. Fifty-three per cent said they had a defined strategy and were actively building for it, while another 37% were exploring the area without a defined strategy.
AI agents are already being used across an average of 3.6 marketing functions per organisation. Search, SEO and AEO management was the most common use case, reported by 48% of respondents, followed by customer service and post-purchase engagement (46%) and campaign personalisation and segmentation (44%).
However, organisations differ over how much autonomy to give AI systems.
Forty-three per cent said they would allow agents to operate autonomously across most marketing tasks, subject to periodic review. Thirty-four per cent would restrict them to low-stakes decisions, while 21% require human approval before an agent action can be executed.
“How do I enforce workflow to ensure that anything that has been touched or used by AI has been reviewed by a human before it goes out the door?” said Sara Sullivan, SVP of Solution Engineering at Contentful.
The report also highlights a gap between smaller and larger marketing departments. Twenty-eight per cent of small departments would extend agent autonomy across most tasks, compared with 45% of larger teams. Similarly, 36% of small departments have a defined agentic-commerce strategy, versus 56% of larger departments.
Human judgement remains a concern
While AI is taking on more operational work, marketers remain cautious about using it for decisions requiring human judgement.
Creative ideas and strategic decisions were the areas respondents trusted AI least, both at 44%. Brand voice followed at 43%, customer-facing copy at 41% and performance insights at 40%.
At the same time, 96% of marketing teams reported at least one structural change during the past year because of AI. Forty-six per cent had incorporated automated systems or “digital workers” into workflows, while 43% had added AI or automation-focused roles. More than half (53%) had increased AI training or enablement.
The findings also point to potential changes in the marketing talent pipeline. Twenty-one per cent of organisations had eliminated certain roles because of AI, while 25% had reduced or paused entry-level hiring.
“To write effectively, to communicate effectively, you need to know what your objective is, and you need to be able to measure its impact… not just say, are there em dashes, but rather, does this prove my objective? Did more people buy my product? Did more people share this article?” said Gabriel Dillon, Go-to-Market Lead, Personalisation at Contentful.
Brands face a new visibility challenge
The report suggests AI is also changing the way brands need to think about visibility.
Eighty-five per cent of marketing leaders believe AI summarisation will make most brands sound the same, while 95% agree that brands with strong, well-codified identities will widen their lead in an AI-mediated environment.
Asked what will distinguish brands once AI systems are doing more of the summarisation, 37% of executives pointed to the structure and clarity of content. Twenty per cent cited third-party validation and authority, while 17% selected the content itself.
“Content needs to be structured and well organised. That means that there’s a declarative understanding of what the content is and what it does, so that agents can then access that information and know not just what are the words within the content, but how that content can be applied,” said Dillon.
AEO and GEO are already receiving greater priority than SEO among respondents, at 55% versus 45%.
The research found that maintaining consistent brand information across third-party sources and creating dedicated FAQ or Q&A pages were the most commonly prioritised optimisation activities, at 34% each. Schema markup followed at 31%, while 29% prioritised an MCP server or similar agent-facing infrastructure.
Measuring how AI represents brands
Despite the growing focus on AI visibility, measurement remains inconsistent.
Eighty-three per cent of marketing leaders said accurately appearing in AI systems was a top or high priority for the next 12 months. Yet only 55% said their teams regularly measure how they appear on the agentic web.
The report refers to changes in how AI systems describe, compare or recommend a brand as “representation drift”.
Thirty-three per cent of respondents track recurring buyer questions to identify content gaps, 31% have built or acquired tools to measure agent visibility, and 24% audit brand mentions in AI-generated outputs.
“The question is not whether the data that it gave me was accurate… The question is whether it is convincing enough for it to make a decision. And in this case, one manufacturer got a sale, one manufacturer didn’t get a sale,” said Charlie Bell, Senior Director, Solutions Engineering at Contentful.
The example illustrates a wider issue identified by the report: brands may not know when an AI system has influenced a customer’s decision, or why one brand was recommended over another.
Who owns AI visibility?
Responsibility for AI visibility is spread across organisations.
Marketing has primary responsibility at 45% of companies surveyed, followed by dedicated AI, data or innovation teams at 21% and IT or engineering at 14%. Six per cent reported having no clear owner.
New AI-related roles are nevertheless emerging. Half of respondents said an AI or agent strategist role exists or is under consideration, while governance, ethics or disclosure leads were reported by 42%. AEO managers, agent operations orchestrators and AI brand-safety officers each reached 37%.
The report argues that organisations need clearer accountability as AI becomes more involved in how customers discover and evaluate brands.
“One of the ways that we use these agents is to better talk to humans, and to talk to other agents… The task of the marketer has always been, how do we reach our buyer, our consumer? How do we think about the person that we’re trying to talk to?” said Dillon.
The research concludes that the challenge for marketing teams is increasingly operational rather than simply technological: deciding what AI systems can do, where human judgement remains necessary, how brands are represented in AI-generated answers, how those representations are measured and who is responsible for correcting them.
