The Agentic Marketing Revolution

Most C-suite conversations about agentic AI are answering the wrong question.

The question they are asking is: how do we use AI agents to make our marketing faster, cheaper, and more personalised? It is a reasonable question. It is also the less consequential one.

The question that should be keeping CMOs awake is this: what happens when AI agents become the buyer? When a consumer’s agent evaluates your brand, compares your offer, and executes the purchasing decision, without a human ever visiting your website, seeing your campaign, or experiencing your brand story?

Brands are already seeing web traffic decline by five, ten, even thirty per cent as AI engines absorb more of the customer journey. Fortune 500 CMOs and CEOs are now identifying AI visibility as their number one strategic priority. Yet most organisations are still investing as though the only disruption is operational: faster campaigns, better content, leaner teams. They are preparing for the last war.

This briefing draws on interviews with nine senior practitioners to separate the agentic revolution into its two distinct frontiers, explain why conflating them leads to misallocated investment, and map the strategic decisions that will determine which organisations lead, and which face strategic irrelevance.

At Salesforce, Martin Kihn describes a cascading agent architecture for campaign creation: one agent develops the brief, a second pulls the audience segment, a third builds the media plan, a fourth handles optimisation. Together they replicate what previously required weeks of cross-functional coordination.

At Klaviyo, Andrew Bialecki has launched what he calls an autonomous CRM, a system that audits a business, identifies campaign opportunities the operator had not considered, generates briefs, and converts them into live marketing constructs. The marketer becomes what Bialecki calls the editor-in-chief: reviewing work, not originating it.

Timo Weis, Global Head of Growth at Infosys, captures the magnitude in practical terms. Campaign duplication that took weeks ten to fifteen years ago can now be automated in minutes. The review, not the creation, is now what takes the longest.

How Agentic AI Changes Campaign Workflows

Martin Kihn, who spent years as a Gartner analyst covering marketing clouds, applies the hype cycle framework directly to the current moment. Customer data platforms went through the same cycle in 2018, initially overhyped, then plunged into the trough of disillusionment as organisations discovered implementation was harder than anticipated. He sees the same trajectory forming around agents.

Dennis Buchheim at Snowflake applies an eighty-twenty rule: not everything will happen as quickly as the most enthusiastic projections suggest. But his five-year outlook is stark. By 2030, the industry will be almost unrecognisable. He describes the pace of change as faster than the shift to programmatic, which was itself the most disruptive transition advertising had experienced.

Dennis Claus, Strategy Director at Apply Digital, anchors this in practical leadership terms. His counsel: think less in terms of milestones and more in terms of momentum. The organisations that win will not be those that made the boldest bet, but those that built the most sustainable flywheel.

Tony Marlow frames it most directly. In a world where consumers delegate purchasing decisions to AI agents, brands will need to market not only to humans but to the agents themselves. A consumer tells their agent to book a trip, specifying a preference for a particular airline but adding a cost threshold. The agent evaluates options, and the brand that has structured its information to be discoverable and preferred by the agent wins the transaction.

The implications are structural. If AI agents mediate an increasing share of consumer decisions, the traditional marketing funnel, built around capturing human attention, shaping perception, and driving conversion through branded touchpoints, loses relevance at the exact point where the decision-maker is no longer human.

The Machine-Mediated Journey

Lindsay Boyajian Hagan at Conductor reports that Fortune 500 CMOs now identify AI visibility as their number one strategic priority. The average prompt to an AI search engine is now twenty-three words, users are specific, conversational, and expect a direct answer. AI engines own the synthesis. They decide which brands get cited, which products get recommended, which companies get surfaced.

Kihn states it plainly: Agentforce is built on Data Cloud. You need the data foundation first, and then agents on top. Bialecki’s entire product philosophy at Klaviyo rests on the same principle, his company spent its first years building a database that combined the real-time responsiveness of transactional systems with the analytical depth of data warehouses, because neither alone could power the autonomous CRM he envisioned.

Buchheim at Snowflake advocates for composability: rather than monolithic marketing suites, businesses should discover and connect the data sources and applications that best serve their needs, with agents orchestrating across them. The composable approach will drive the convergence of ad tech and martech that the industry has discussed for years but failed to deliver.

Mitali Israni, Senior Marketing Director at Pantheon, offers the practitioner’s perspective. She describes AI as helping teams track performance and allocate budget with far greater precision through predictive analytics, identifying which accounts are genuinely in-market rather than wasting spend on those who are not ready to buy.

This is the least glamorous section of this briefing and the most important. Data unification is expensive, slow, and invisible to the board until it fails. Leaders who fund the agent layer before the data layer are building on sand.

The Skill Erosion Risk

Martin Kihn raises a question most AI advocates avoid. If teams delegate all their writing to language models, their writing will deteriorate over time. If they delegate all research, their capacity for original insight will atrophy. He also identifies a subtler risk: the erosion of internal brand capital. When colleagues receive communications that are obviously machine-generated, it signals a lack of care.

Timo Weis at Infosys extends this concern to the pipeline of future talent. Many tasks now handled by AI, data analysis, creative development, campaign structuring, are precisely the activities junior marketers need to learn the fundamentals of the discipline. His caution: we can’t outsource our brains to AI.

The Emotional Territory Agents Cannot Occupy

AI agents excel at evaluating structured attributes, price, availability, specifications. They are structurally incapable of experiencing what Darren D’Altorio, CEO of Wpromote, calls the emotional layer that breaks through: the sense of inspiration, belonging, and identity that transforms a transaction into a relationship.

Ruslan Tovbulatov, VP Marketing at ServiceNow, crystallises this into a principle: brand becomes the most important moat. Not the logo or the colour palette, but the relationship a company has with its customers. In a world where AI can replicate everything else, that relationship is the one asset that cannot be vibe-coded.

Lindsay Boyajian Hagan frames the positive case. Human creativity is resonating more than ever precisely because AI can produce everything else. Events are selling out. People are craving human connection. The scarcer the craft, the higher its value.

1. Separate the two frontiers in your investment thesis

Internal agentic AI delivers returns now and should be funded as operational transformation. External agentic AI, AEO, machine-readable brand architecture, is a medium-term strategic bet that requires research investment and cross-functional ownership. Budget them separately. Govern them separately. Measure them against different horizons.

2. Protect human craft as a strategic asset

As AI-generated content proliferates, the capacity for original thinking, genuine brand voice, and human communication becomes scarcer and more valuable. Actively develop these capabilities. Hire for judgment, invest in craft development, and create structures that keep humans practising the skills that matter.

3. Prepare your brand for machine readability, now

Audit your AI visibility. Build topical authority around the categories you want to own. Invest in AEO as a strategic capability, not a tactical experiment. This is infrastructure, not a campaign. If an AI agent cannot find, parse, and compare your offering, you do not exist in the machine-mediated journey.

4. Fund the data layer before the agent layer

Every practitioner identifies data quality, unification, and governance as the binding constraint on agentic effectiveness. Agents built on fragmented or ungoverned data produce outputs that erode brand trust faster than manual processes ever could. If you cannot answer “is our data ready for agents?” with confidence, that is your first investment.

5. Resist the trough, and sequence your deployment

Expect short-term friction as agents are tested in workflows. Kihn’s distinction between employee-facing agents (lower risk, immediate productivity gains) and customer-facing agents (higher risk, brand exposure) provides a practical sequencing framework. Build confidence and governance internally before deploying agents into customer interactions.

The briefing in your hands is the diagnostic. The 90 days that follow are the test.

Dennis Buchheim’s five-year prediction is the most useful frame for calibration. When pressed for a bold forecast, his answer is four words: I think our agents are talking. Andrew Bialecki sees the same future from the infrastructure side: a world in which AI search engines query the agents that represent businesses directly, rather than crawling static website content.

The question for every C-suite leader is no longer whether this happens. It is whether your organisation is building for both frontiers, or betting everything on the one that feels familiar.

Strategic irrelevance does not arrive with a crisis. It arrives with a series of quarterly results that look slightly worse than the last, while the organisations that made different choices compound their advantage.

The Voices Behind This Briefing

This briefing is distilled from interviews conducted on The Business of Marketing podcast.

Additional perspectives referenced: Timo Weis (Infosys), Dennis Claus (Apply Digital), Ruslan Tovbulatov (ServiceNow), Salomé Imedashvili (Salesforce Partnerships).