What should marketers do as AI agents buy more?

Tim Soulo

CMO · Ahrefs

Continuous learning

Tim Soulo's answer is continuous learning: AI can generate endless ideas, but experience is needed to know which are good and when to stop, and experience comes from doing things and learning from others.

The answer starts at 54:10.

Daniel Murray

Founder and Creator · The Marketing Millennials

Human‑in‑the‑loop

Daniel Murray says marketers should keep a human in the loop, arguing you need a human to audit what an AI agent is doing.

The answer starts at 24:36.

Duarte Garrido

Co-Founder & Co-CEO · DOJO AI

Creative ideation

Duarte Garrido says marketers should prioritize creative ideation, since AI is built on logic and not suited to creativity, which comes from human experience, and logic alone won't differentiate a brand.

The answer starts at 29:30.

Gary Lombardo

Adjunct Professor

Small continuous experiments

Gary Lombardo says his team builds a standing rhythm of small experiments—testing a channel, checking results, adjusting—rather than waiting for a big annual plan to be right.

Rustom Dastoor

Executive Vice President, Head of Marketing and Communications for the Americas · Mastercard

Become an intelligence function

Rustom Dastoor says marketing should become an intelligence function, using data and AI as an early-warning system that influences product development and pricing, rather than remaining just a creative function.

The answer starts at 20:01.

Cathy McPhillips

Chief Marketing Officer · Smarterx

Internalize AI capability

Cathy McPhillips says marketers should internalize AI capability, arguing they need to maintain the IP and know-how within the organization and that it is a fundamental skill to build internally.

The answer starts at 16:20.

Paul Roehrig

Chief Strategy and Marketing Officer · Ascendion

Focus on storytelling

Paul Roehrig says the nuts and bolts of marketing come down to having a good story, since if you don't have a good story the tactics are much less relevant.

Marvin Chow

VP of Consumer and AI Marketing · Google

Systems thinking

Marvin Chow says marketers should develop systems thinking — looking across an enterprise to spot frictional pain points and treat them as problems to solve — because it will be a valuable skill as tools move toward agents and systems.

KM

Khurrum Malik

VP of Business & Product Marketing · Walmart Connect

Prioritize human trust

Khurrum Malik says AI can't replace trust, human connection and human accountability, and that he's leaning into relationships instead.

Mark Abramowitz

Chief Marketing Officer · Dataiku

Be persona‑specific

Mark Abramowitz says marketers should be very persona-specific in their positioning and messaging, since the broad buying committee makes it even more important to be precise about who they are talking to about what.

Dawn Dawson

Chief Marketing Officer · Mortgage Solutions Financial

One‑to‑one experiential experiences

Dawn Dawson says marketers should create true experiences and make them one-to-one as much as possible, arguing that as AI increases efficiency it also isolates people, who need human connection.

Lucinda Barlow

Senior Director, Head of International Marketing (APAC, EMEA, Latin America) · Uber

Bring a creative spirit

Lucinda Barlow says to bring a creative spirit—meaning curiosity, openness and a willingness to embrace smart, sometimes scary risk-taking—to achieve great impact and growth.

Rachel Thornton

CMO, Enterprise · Adobe

Educate agents

Rachel Thornton says marketers should treat agents as a new audience and educate them, making them aware of what the company offers, since agents may help evaluate or decide purchases.

Jonathan Kerr (JK)

Chief Growth Officer · Auto & General Australia

Become AI native

Jonathan Kerr says marketing teams need to become AI native, arguing they shouldn't feel threatened and should instead aim to be orchestra conductors who pull the band together.

Mark Ritson

Founder · Minimba

Embrace synthetic data

Mark Ritson says B2B marketers need to embrace synthetic data, saying it is now more accurate than surveys and can do conjoint pricing, survey design and segmentation.