Basis has written a readiness guide for agencies that want an agent to spend their clients’ money. In Is Your Agency Ready for Agentic Advertising?, the argument is that the tool is the wrong first question: “what actually determines success is whether an agency’s data infrastructure can support an agent that plans, decides, and executes on its own.” The warning is specific. “An agent pointed at fragmented, stale, or un-auditable data won’t recognize anything is wrong, and it will still act, executing the errors at machine speed.” Readiness comes down to four dimensions, data centralization, workflow integration, system-of-record integrity, and governance and auditability, each with a not-ready, partially-ready and ready column.
The figures are a mix of third-party and vendor, and the post labels most of them. Two-thirds of ad buyers call agentic buying an increased focus and 84 percent name media planning recommendations as a use case, from the IAB’s 2026 outlook study. From Basis’s own agency report, 36.8 percent of agencies now manage ten or more adtech tools, more than double two years ago, with inefficient processes at 44.1 percent and siloed systems at 40.4 percent as the top complaints. The Forrester Total Economic Impact study Basis commissioned found consolidation cut manual steps by 40 percent and media operations time by 43 percent, which is a vendor-sponsored number about the vendor’s platform. What survives independent of the marketing is the governance question the post poses: “If a client asked why an agent moved their budget last week, could the agency show them the reasoning and the trail behind it?”
The catalog can ask the same question of Basis. The Basis provider page describes REST APIs for campaign management, audience targeting, publisher management, reporting and omnichannel buying, and the record carries one API page, the Basis Analytics API. The agentic access profile maps 31 operations. On the Agent Readiness score, spec presence, agentic access, auth clarity, error semantics, examples, rate limit signals, agent skills, delegated identity and dynamic client registration are all lit, which is a better identity layer than most of the providers this series covers. The MCP server and the event surface are unlit, so an agent gets a contract but no push channel to learn that a budget moved.
The Kin Score is 54.2, developing band, with discoverability at 66.1, contract quality at 60.2, access clarity at 57.9, developer ergonomics at 44.6, contract governance at 41.7 and operational transparency at 31.6. The Agent Readiness score is 40.8, agent-ready. The post’s fourth dimension is governance and auditability, the ability to trace what an agent did and why. Operational transparency is the facet that measures the nearest thing on a public record, and it is the lowest of the six. Basis says readiness is an infrastructure question before it is an AI question. The record agrees, and says the infrastructure an outside agent can see is one analytics API.