run it from your own AI
an AI marketing agent that checks its own work.
It reads your ads, analytics and search data, spots the opportunity, ships the change once you approve, and comes back three to seven days later to say whether it worked. In the Cogny app, or inside Claude, ChatGPT, Cursor or Codex.
or from your own AI
claude mcp add --transport http cogny https://app.cogny.com/mcpwhat an AI marketing agent does
An AI marketing agent is an assistant that can do marketing work, not just talk about it. It needs three things a chatbot doesn't have: access to your live data, permission to change things in your tools, and a way to find out afterwards whether the change helped.
Cogny provides all three. It connects to 72 marketing integrations, gives the agent read and write tools for them, and turns every change into a ticket that is measured in your own data and closed with a verdict. Misses are recorded too, so the agent's next proposal starts from what actually happened.
Not “here's what I'd do”. Done, measured, and written down.
marketing doesn't stop. your team does.
Too much data, too little time
Nobody reads every campaign, query and funnel step every week. Signals get buried.
Opportunities expire
A rising query or a tired creative is worth acting on this week, not next month.
Testing is slow
Research, setup and checking eat the time that should go to deciding.
Nobody follows up
Changes ship and nobody checks a week later whether they worked.
connect once. the agent does the rounds.
- 01
Connect
OAuth into your ad platforms, analytics, search, CMS and email.
- 02
Spot
Ask, or open the start page: a graph per KPI, last 30 days against the 30 before.
- 03
Act
One proposed change, shipped in the tool once approved.
- 04
Validate
Before against after, 3 days for ads and 7 for content. A verdict, with the numbers.
evidence, not a status update
Example rows, not customer data.
a marketing teammate that shows its work
In the AI you already use
Add Cogny to Claude, ChatGPT, Cursor or Codex with one line, then authorise in the browser.
Or in the Cogny app
Chat and a voice agent, a start page with a graph per KPI, and a daily update on paid plans.
Every ticket measured
Tickets close as worked, partial, didn't work or inconclusive.
It knows your brand
Brand & Context is built from your website at signup, free on every plan, and it never overwrites what you wrote.
what it doesn't do
- It doesn't act behind your back: in the app write tools ask first, and in your own AI its approval prompt is the gate.
- It doesn't watch every metric around the clock. It works when asked, on the schedules you set, and in the daily update on paid plans.
- It doesn't grade itself on platform metrics. Verdicts are read from your own connected data.
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 →
an agent that keeps the ledger.