run it from your own AI
agentic marketing, with a record of every move.
An AI agent that researches your data, proposes the next change, makes it in the real tool once you approve, and measures what it did. Run it from the Cogny app or from the AI you already use. Cogny keeps the ledger.
or from your own AI
claude mcp add --transport http cogny https://app.cogny.com/mcpwhat agentic marketing is
Agentic marketing is marketing work done by AI agents that can act, not only answer. A chatbot tells you what it would do. An agent connected to your stack reads the live numbers, decides on one next step, and makes the change in Google Ads, Meta, your CMS or your email tool.
Cogny is that connection, and the memory around it. It gives the agent 72 integrations with read and write tools, and it gives you the record: every write logged with who authorised it, every change turned into a ticket (from your own AI, launches automatically and other changes when you register them), every ticket measured in your own data and closed with a verdict.
The agent can be Cogny's own, in chat or by voice in the app, or yours: Claude, ChatGPT, Cursor or Codex over MCP. Either way a person stays in the loop. In the Cogny app write tools ask first by default; from your own AI, its approval prompt is the gate. What you set as blocked is refused.
Agents propose. You approve. The ledger remembers.
why traditional marketing hits a ceiling
People are the bottleneck
Every change needs someone to research it, make it and check it. Output grows with headcount.
Strategy dies in execution
The plan is clear in the deck. The changes never quite all ship.
Optimising after the fact
Problems are found in the monthly review, weeks after they started.
Agents without memory
A chat that made a change last week doesn't know whether it worked. Nothing is measured.
you set the direction. the agent does the legwork.
- 01
Ask
A goal or a question, in the Cogny app or your own AI.
- 02
Research
The agent reads ads, analytics, search, email and site data together.
- 03
Propose
One concrete change, with the number it should move.
- 04
Approve and ship
Made in the real tool once you say yes. Logged.
- 05
Verdict
Measured 3 or 7 days later. The next step starts from it.
what an agent's week looks like on the ledger
Example rows, not customer data.
a marketing operation that scales without the headcount
An agent with real tools
72 integrations with read and write tools, from Google Ads to WordPress to Klaviyo.
Your AI or ours
Claude, ChatGPT, Cursor, Codex, or the Cogny app with chat and voice.
Human-in-the-loop by default
Write tools ask first in the app; per-tool settings; blocked tools are refused.
Loops for what works
A finished chat becomes a Loop on a schedule (Solo, Cogny AI, Cogny Pro). Each run is a measured ticket.
what it doesn't do
- It doesn't run your marketing unattended. Writes are approved, or pre-granted per tool for a Loop you set up.
- It doesn't learn in secret. What it knows is your data, your brand context and the record of past verdicts.
- It doesn't replace strategy. You choose the goal; the agent does the research, the change and the check.
the record.
Published case studies, each one linked.
what it costs.
- Free $015 MCP calls a month and 50 one-time chat credits, no card
- Solo $9unmetered in your own AI, fair use
- Cogny AI $18150 credits
- Cogny Pro $901,500 credits
- Cloud $4995,000 credits, for teams
Per month. Solo, Cogny AI and Cogny Pro have one seat each. Every change is recorded on every plan; Solo measures 3 a month, Cogny AI and Pro measure every one and add monthly credits. Teams need Cloud. Compare plans →
agents that keep score.