Sentient marketing is customer engagement built on four traits: personalized to the individual, run as two way dialogue instead of one way broadcast, delivered in real time, and executed at full scale through automation. Strategist Mohanbir Sawhney and Teradata executive Oliver Ratzesberger defined the term in January 2019. Salesforce found in February 2026 that just over half of marketers can reliably respond to a customer message, and Supermetrics found only 6% have fully embedded AI into their workflow. The definition has not moved in seven years. The distance between the definition and the practice is the entire story.
The term has a named origin and a specific meaning
Sentient marketing did not emerge from a conference hashtag. It came from a book. Mohanbir Sawhney, a strategy professor at Northwestern’s Kellogg School of Management, and Oliver Ratzesberger, then chief product and technology officer at Teradata, coauthored The Sentient Enterprise, arguing that a company should sense and respond to its environment the way a living organism does, adjusting its next move continuously instead of waiting for the next quarterly plan. They laid out the marketing application of that idea in a Forbes CMO Network piece published January 14, 2019, titled “Sentient Marketing: Customer Engagement At The Speed Of Data.” Their examples were autonomous vehicles and algorithmic trading systems, both of which sense conditions and act in milliseconds with no human in the loop.
Applied to marketing, Sawhney and Ratzesberger described four traits a company needs before it can call its engagement sentient. Personalized, meaning treatment of the individual rather than the segment. Two way, meaning the customer can respond and be heard, not just receive a broadcast. Real time, meaning the gap between a signal and a response shrinks toward zero. And at scale, meaning automation carries the weight a human team physically cannot.
Worth naming plainly: several companies and one later book have also used “sentient marketing” as a brand name, generally to describe emotionally attuned or socially conscious brand strategy rather than Sawhney and Ratzesberger’s data and automation framework. This piece uses the original, most cited definition, the one Forbes published and Kellogg’s own faculty page still credits to Sawhney.
Seven years on, the vision outran the infrastructure
The world Sawhney and Ratzesberger described in 2019 has since been hit by two forces they did not fully anticipate: large language models and, more recently, autonomous AI agents that browse, compare and buy on a customer’s behalf. Salesforce’s tenth State of Marketing report, drawn from a double anonymous survey of 4,450 marketing decision makers between October and November 2025, found that 85% of marketers believe AI is reshaping SEO strategy and 88% are already optimizing content for AI generated answers rather than for the traditional ten blue links. Real time, in other words, now has to account for an AI system reading and summarizing a brand before a human ever sees the page. That is a harder, faster version of the real time Sawhney originally meant.
The same report found that just over half of marketers can reliably reply to a customer’s email or text message, and 69% say they struggle to respond to customer interactions promptly. Eighty three percent believe customers now expect an ongoing two way conversation rather than a one way campaign. The expectation and the capability are not the same line on the chart.
Supermetrics surveyed 435 marketing leaders for its 2026 Marketing Data Report and found the constraint sits underneath the tools, not inside them. Eighty nine percent said pressure from the C-suite or board to adopt AI has increased, yet only 6% have fully embedded AI into their workflow. Fifty two percent said no one is clearly accountable for their organization’s data strategy. Just 24% of marketers say they have achieved personalization at scale. And MIT’s Project NANDA, which studied more than 300 enterprise generative AI deployments, found 95% of pilots produced no measurable effect on profit or loss, largely because the tools were bolted onto broken workflows instead of replacing them. None of this contradicts Sawhney’s 2019 thesis. It confirms it. He described a gap between vision and execution. Three separate, independently run 2025 and 2026 studies just measured that exact gap and found it still open.
Sentient marketing was never a tool you buy. It is a description of what a company becomes once its data, its automation and its judgment finally operate as one system, which happens to be the exact part most AI roadmaps skip.
Aware, Adaptive, Alive is what sentient marketing looks like once it is built
We did not invent the underlying idea. We built a practical operating model around it, because Sawhney and Ratzesberger were writing for Jio, Sephora and Microsoft, companies with data science teams larger than most of our clients’ entire staff. A mid-market company does not get there by buying the same platform Sephora bought. It gets there by sequencing the same four pillars in an order a lean team can actually execute.
Aware maps to sense: the unglamorous work of unifying data that currently lives in three or four disconnected systems, so the company can see a customer instead of guessing. Adaptive maps to decide and act together: building the two way, real time response Salesforce found only about half of companies can currently deliver. Alive maps to scale: letting automation carry the volume once judgment has been proven at a small scale first, the same discipline we describe in Machines For Scale, Never For Soul. None of this replaces Sawhney’s original definition. It operationalizes it for a company that has a marketing team of three, not three hundred.
Unify before you buy
Connect the two or three systems that already hold customer data before adding a fourth AI tool that cannot see any of them.
Turn on real replies
Fix the ability to respond to an inbound email or text within the hour before investing further in outbound personalization.
Prove one segment first
Personalize one high value segment completely before promising personalization at scale to the whole list.
Name the data owner
Assign one person accountable for data strategy. Fifty two percent of marketers currently have no one in that seat.
Write for the AI reader
Optimize the pages that answer real questions for AI generated summaries, not only for the traditional search results page.
The part nobody wants to hear
Sentient marketing stayed aspirational for seven years for the same reason most good strategy stays aspirational: the work that makes it real is unglamorous and belongs to no one’s job description. Unifying three data sources does not make a deck. Naming a data owner does not launch a campaign. It is Tuesday afternoon work, and Tuesday afternoon work loses to whatever is loudest that week, which is precisely the pattern this publication keeps finding underneath nearly every strategy that stalls.
Sawhney and Ratzesberger were not describing a future that would arrive on its own. They were describing a company that chooses, deliberately and repeatedly, to sense before it guesses and to respond before it broadcasts. Seven years and one generative AI boom later, that is still a choice almost no one has fully made. The 6% who have are not smarter. They just did the unglamorous part first.