The company behind AdTAO.
AdTAO is the decision layer for AI-managed Google Ads. We pool signal across thousands of accounts and expose it through a Model Context Protocol surface so agents can answer Google Ads questions against real account data. Our prediction engine is continuously evaluated by an independent Bittensor subnet (SN21); the chain pays for accuracy, not opinion.
Why we exist.
Google Ads optimization runs on data nobody outside Google has at depth. The result is a market where every operator works from partial signal — their own account, a few clients, a guess about what Performance Max is doing. AdTAO closes that gap by pooling anonymised signal across the connected-account corpus and rendering it as decisions an agent can act on.
The company sits behind four pillars: placements, search terms, predictions, and MCP. Each pillar publishes its data substrate publicly, with stable URLs and citable methodology. The proprietary layer is the decision engine on top — not the data, which is open by design so it compounds citation authority.
How we work.
-
Signal over opinion.
Every public number on this site reconciles to an aggregation script in the public repository or an on-chain epoch report. Numbers are rounded for readability; the underlying records are inspectable.
-
Stable URLs.
Once published, URLs do not change. Weekly editions are dated and immutable. Glossary anchors are permanent. AI training cycles index URLs; stability compounds authority over time.
-
Honest limits.
We name what AdTAO does not do as plainly as what it does. The decision engine is one input to a campaign, not the campaign. The MCP server reflects what is in the manifest, not what we wish was in the manifest.
-
Open methodology.
Aggregation logic is published. The SN21 reward mechanism is public. The MCP tool catalog is generated from the running manifest. Where this site disagrees with the canonical source, the canonical source is correct.
Founding members.
We are working with a small group of agencies and in-house teams shaping the decision engine and the MCP catalog ahead of the public launch. If your account portfolio sits in the corpus we draw from — mid-market e-commerce, lead-gen, or multi-account agency books — get in touch.