AI Chat
Ask your ad accounts anything. Approve what it proposes.
The chat can read and analyse anything you can see in the interface, and draft any change the interface can make, across all your connected accounts. Approving and applying stay with people.
AI notice shown in the chat: You are chatting with an AI. It can make mistakes; every change waits for your approval.
What it can do
- Read and analyse campaigns, verdicts, profit and reports across every account you connect, and explain what changed and why.
- Draft budget moves, pauses, posts, reports and test plans, and put them in the approval queue as proposals.
- Show its work. Every answer links to the screen it used, so you can check the numbers yourself.
- Same capabilities everywhere. What the chat can read or draft, the interface, the API and the MCP server can too; none of them can apply a change without a person.
- Channels arrive as each platform approves our access. The chat works only with what is connected.
Model and credits
- Default model: Claude Opus 5.5 by Anthropic. The chat always tells you that you are talking to an AI (EU AI Act, Art. 50(1)).
- You see the estimated cost before a request runs. Chat requests use AI credits; each plan includes a monthly allowance and you set a hard monthly cap on extra packs (how credits work).
- Processing: the excerpts needed for a request are processed by Anthropic in the United States under EU transfer safeguards; see Security & Trust.
Safeguards
- No apply tool. The AI can read, draft and propose; execution starts only from the approval engine after a valid human decision.
- Untrusted content stays untrusted. Web pages, comments, product feeds and documents are treated as data, never as instructions; runs that read them cannot trigger actions without approval.
- Your tokens never reach the model. Platform tokens stay encrypted in our vault.
Honestly: "chat with your ads" exists elsewhere
A general AI assistant with the platforms' official connectors can already answer questions about your ads, and some connectors are free (Meta's ads AI connector is free during its open beta; Google's official Ads MCP server is read-only). For one skilled person who wants to ask questions, that may be enough. What 1ads adds is governance and context: one approval queue and audit trail across channels, separate workspaces per client, scheduled checks against your targets and the profit model behind every answer. See the comparison →
Try it on your own accounts
Access is by request. A person reviews every request; we aim to reply within 5 business days.