Gartner just put a number on how fast marketing is supposed to change. The research firm forecasts that 60 percent of brands will use agentic AI to run streamlined, one to one customer interactions by 2028, with agents processing orders, flagging service problems before a customer notices, and tailoring financial recommendations in real time. Emily Weiss, a senior principal researcher in Gartner’s marketing practice, told the firm this marks the end of channel based marketing as it has worked for the past two decades. That is a confident claim about infrastructure that, by most other measures, is not built yet.
Right now, full automation is rare, not standard. A compilation of 2026 marketing automation surveys puts the share of marketers running a fully automated, end to end customer journey at around 9 to 10 percent, even though 96 percent of marketing teams have adopted or plan to adopt some automation platform. The same compiled data shows 44 percent still automate only part of a journey by hand, stitching email, retargeting, and lifecycle rules together whenever a campaign changes. This figure comes from aggregated survey roundups rather than one named study, so treat the exact percentage as directional rather than precise.
A more carefully sourced study shows why that gap persists. The AI Marketing Maturity Study 2026, run by the Marketing and Media Alliance (MMA) with EY and Mobavenue through interviews with more than 60 senior marketing leaders across CPG, BFSI, technology, and real estate, found that content and creative generation is the most mature AI use case in marketing, and even there 65 percent of teams still rely on stand alone tools that are not connected to the rest of their stack. Customer and sales journeys scored a little better, with 54 percent reporting full stack implementation, but a meaningful share of those still depend on an outside agency to run them day to day.
That is the part the 2028 forecasts leave out. An autonomous agent making a one to one decision needs a clean handoff between the system that knows the customer, the system that holds inventory or policy rules, and the system that acts. Most marketing stacks today, per the MMA and EY numbers, do not have that. They have capable tools sitting next to each other instead of inside each other. Putting an agent on top of a disconnected stack does not remove the disconnection, it just moves the point of failure from a person who notices a broken handoff to an agent that acts on bad information with full confidence. Gartner may still be right about the direction. The timeline assumes a wiring job most marketing departments have not started.
For a marketing leader evaluating an agentic AI pitch this quarter, the useful question is not whether the vendor’s agent is good. It is whether the systems that agent would need to talk to, CRM, order management, policy and compliance, already share clean data today without a human stitching it together by hand. The MMA and EY study found 85 percent of teams reported medium to high business impact from AI, but that was measured on narrow, supervised use cases such as content drafts and propensity scores, not on autonomous one to one journeys running without review. Confusing one kind of success for the other is how a 2028 forecast turns into a 2026 budget line that does not deliver.
Sources referenced in reporting:
Digital Commerce 360: Gartner predicts 60 percent of brands will use agentic AI by 2028
MMA and EY: AI Marketing Maturity Study 2026
Thunderbit: Marketing Automation in 2026, 45 Stats and Insights