Every Change Gets a Verdict: The New Cogny
Today cogny.com looks different, and it says one thing first: every change gets a verdict, misses included.
This post is also in Swedish.
Most marketing tools stop at the report. Your AI assistant can now change your ads, your site and your CRM directly. Nobody checks whether the change worked. Cogny does.
What Cogny does
Experiment across your whole stack. Change something in one tool and measure it in another: ads, analytics, search, email, CRM, your site and your code. Every integration Cogny connects to, and every read and write tool in it, is listed in public.
From raw data to a shipped change, then a verdict. Ask a question about your numbers. Cogny proposes one concrete next step. You approve it, it ships in the tool, and it becomes a ticket. Cogny measures the ticket in your own connected data, 3 days later for ads and 7 days later for content and SEO, and closes it with one of four verdicts: worked, partial, didn't work or inconclusive.
Run it from your own AI. Cogny keeps the ledger. Use Claude, ChatGPT, Cursor, Codex or the Cogny app. Every write is logged with the tool, the time and the result. When your AI launches something through Cogny, such as a campaign, an ad or a post, Cogny opens a ticket for it by itself. For other changes, your AI offers to register them. On Solo, Cogny AI and Cogny Pro, a finished chat can become a Loop that runs on a schedule, and each run is its own measured ticket.
On every plan
Verdicts are not a Cloud feature. Every plan gets measured tickets:
| Plan | Price | What's included |
|---|---|---|
| Free | $0 | 15 MCP calls a month, 50 one-time chat credits, no card |
| Solo | $9/month | Unmetered MCP from your own AI (fair use) |
| Cogny AI | $18/month | Solo plus 150 credits a month |
| Cogny Pro | $90/month | Solo plus 1,500 credits a month |
| Cloud | $499/month | 5,000 credits a month, team workspace, per-member access, scheduled reports, pull requests to your repo |
Measurement runs even when a workspace has no credits left. Solo, Cogny AI and Cogny Pro have the same features; only the credits differ. Compare plans.
We publish our own record
A promise about misses needs misses behind it. So we're publishing ours.
- The verdict ledger. Cogny markets itself with Cogny. We counted our own measured marketing changes this year: 167 tickets, 70 worked, 78 partial and 19 didn't work. The page has the method and the counts by kind of change.
- The AI visibility index. We asked ChatGPT, Claude, Gemini and Perplexity 50 questions a marketer asks when choosing tools. Cogny came up in 5 of the 45 that don't name us. That's low, and it's the baseline we'll measure ourselves against.
- The truth ledger and our case studies: a Scandinavian activewear brand where organic search became 50% of orders in six months, and GrowthHackers.se, +271% organic clicks in three weeks.
- Cogny Bench, our benchmark of AI models on marketing-analytics problems.
Everything is collected at cogny.com/research. Journalists will find facts, logos and the boilerplate at cogny.com/press.
Try it
Start free, connect one source and ship one change. A few days later you'll know whether it worked.
In Claude Code:
claude mcp add --transport http cogny https://app.cogny.com/mcp
Then authorise in the browser. Or start in the Cogny app.