# AI Agents Are Making Marketing Decisions On Data No One Has Checked In Years > Source: > Published: 2026-07-01 04:35:46+00:00 Ask anyone in ad tech about data governance, and they’ll talk about the supply side. The industry spent years building verification frameworks for publishers and sellers, proving that the data powering supply-side decisions is what it claims to be. Standards exist, enforcement is maturing and the consensus is clear: If automated systems act on your data, you need to prove it’s accurate. The buy side has no equivalent discipline. And it needs one. AI agents are now making send, suppress and target decisions based on data that nobody has verified in years. **The data underneath the decisions** In most B2B marketing automation environments, the data layer feeding campaign execution was built three to five years ago. Consent records were captured under a regulatory framework that has since evolved. Suppression lists were reconciled against systems that have been deprecated or migrated. And somewhere in there, someone calibrated a lead scoring model against a buyer profile that no longer reflects who actually converts. But none of these issues get flagged since, technically, nothing is wrong with the data. It was all validated, just too long ago. **Why agents change the risk profile** Before AI agents, stale buy-side data was a performance problem. Campaigns underperformed, deliverability drifted, and occasionally the wrong people got the wrong messages. But a human usually reviewed the send list or checked the segment before anything went out, and the worst of it got caught. Agents skip that step entirely. An AI agent making send decisions inside a marketing automation platform acts on whatever the data layer tells it at speed, at scale and with zero instinct to question whether a consent record still means what it says. An agent will suppress an entire segment of high-value prospects based on a rule nobody remembers writing. It’ll send campaigns to contacts whose opt-in intent expired two regulatory cycles ago and keep doing it for weeks until someone notices the pipeline drying up and starts asking questions. That’s the kind of failure that shows up as a deliverability incident, a compliance exposure or a pipeline problem that nobody can diagnose from the dashboard. And it’s the same category of problem the industry spent years solving on the supply side – automated systems acting on unverified data at a scale where human judgment can’t compensate. The difference is that, on the buy side, nobody has named it as a governance problem yet. **What buy-side data governance looks like** The supply-side playbook offers a useful frame: verify, document and create accountability for the data that automated systems act on. For the buy side, the starting point is consent and preference data. Most teams built this infrastructure once and moved on. Pull your opted-in contacts, check when consent was captured and flag anything older than 18 months for reverification against your current processing purposes. Do the same for preference data: If the options in your preference center don’t match the campaign categories your team actually runs today, the data it collects is meaningless. Suppression logic needs the same scrutiny that the industry applies to brand safety rules. Export your suppression rules, trace each one to the campaign or business reason it was created for and kill the ones nobody can explain. If a rule exists and no one on the current team knows why, it’s either protecting you from something important or blocking revenue for no reason. Someone needs to own the data layer that agents depend on, a person who can answer the question, “What data are our agents acting on, and when was it last verified?” On the supply side, data quality has an owner. On the buy side, it sits in a gap between marketing ops, legal and IT. This isn’t sustainable when agents are making decisions on their own. The supply side proved that data governance scales when the industry decides it matters. The buy side needs the same decision before the agents make it for them. *“**Data-Driven Thinking**” is written by members of the media community and contains fresh ideas on the digital revolution in media.* *Follow **Sojourn Solutions** and **AdExchanger** on LinkedIn.*