Expect More in the Long Run
Martin Kihn is SVP Strategy at Salesforce Marketing Cloud and Data Cloud, a role he has held since 2018. A former Gartner Research VP for programmatic and marketing analytics, a seven-time author, and host of the PaleoAdTech podcast, he brings an analyst’s precision and a storyteller’s instinct to the question of where AI in marketing actually goes.
“That is how post-cookie ad execution works. No data foundation, no agents.”
Martin Kihn is SVP Strategy at Salesforce for the Marketing Cloud and Data Cloud businesses. He joined Salesforce in 2018 after serving as Research VP at Gartner, where he led the advisory agenda for programmatic advertising, digital marketing analytics, and the application of AI to marketing. He holds an MBA from Columbia Business School and a Bachelor’s degree from Yale, and is a seven-time author including AgentForce.
Martin’s career began not in technology but in writing. He was a head writer at MTV and worked in screenplay development at Warner Bros before moving into strategy, analytics, and AI at digital marketing agencies. From there he moved to Gartner, where his research on the hype cycle for marketing technology informed the strategic thinking of marketing leaders across the world. He joined Salesforce to move from analyst to operator.
At Salesforce, Martin’s primary thesis is that the data foundation is the prerequisite for everything else: AI models, agents, personalisation, and campaign automation all require a well-organised customer data layer to deliver results. His Agentforce work focuses on building what he calls an agent factory: a platform that allows enterprises to assemble trusted agents, test them, and deploy them safely, distinguishing between employee-facing agents that enhance human work and customer-facing agents that represent the brand.
“The new paradigm is first-party data in Data Cloud, sent to the walled garden to run the campaign.”
“Data foundation first. Agents on top. Without the data, nothing works.”
Martin’s most consistent message about agentive AI is sequencing. The enterprise enthusiasm for agents often runs ahead of the data infrastructure required to make them work. An agent that books meetings, drafts campaign briefs, or handles customer enquiries is only as good as the customer data it draws on. Without a unified, clean, first-party data layer, agents produce outputs that are incoherent, irrelevant, or brand-damaging. The Salesforce Data Cloud is the prerequisite.
“Short-run AI expectations are too high. It is easy to build a single agent.”
Martin applies the Gartner hype cycle framework directly to agentive AI. The CDP category went through exactly this pattern: peak of inflated expectations, followed by a trough of disillusionment, followed by the plateau of productivity. AI agents are in the peak phase. The companies that will win are the ones that do not abandon the investment when the first deployment is harder than expected.
“All of Marc Benioff’s talks take the form of stories. It becomes infinitely relatable.”
Martin’s writing background, which began at MTV and Warner Bros, is the thread he draws on most consistently. The analytical side and the storytelling side are not in tension. They are complementary. Data without narrative is information. Narrative without data is opinion. The combination is what makes a Salesforce keynote or a Gartner research note persuasive rather than merely accurate.
Stay close to the conversation.
New leaders and insights. No noise.