Who is in the crowd around AI agent platforms
We pulled the crowd around eight AI agent platforms in the UK on 20 August 2026. Over 90 days, 4,449 people reacted to or commented on those platforms' public posts.
Then we split them by title. 79 people whose title suggests they sign for software. 1,523 builders and practitioners. Every platform skews the same direction, and the size of the skew is the table below.
This edition's table is a ratio, not reactions per post.
The count: eight platforms, 4,449 reactors, 90 days
Pulled 20 August 2026 in the UK. Eight anchor platforms, a 90 day window, 4,449 raw reactors.
Raw is doing work in that sentence. Before any filter, 4,449 accounts touched these posts. After the title filter two groups came out of it: 79 people whose title suggests they sign, and 1,523 people whose title says they build the thing or run it day to day. The remainder either carried no readable title or did not fit the category.
27 people appear around two or more platforms. For a category with this much noise in it, that is a small number. The crowds are mostly separate audiences that happen to be interested in the same idea.
The ratio table: builders and practitioners per service owner or lead
Sample: 4,449 raw reactors across 8 anchor platforms, UK, 90 days to 20 August 2026. Read each row as the number of builders and practitioners for every one person whose title suggests they sign.
| Platform | Builders and practitioners per service owner or lead |
|---|---|
| Botpress | 11:1 |
| MindStudio | 12.7:1 |
| Relevance | 13.6:1 |
| n8n | 14.8:1 |
| Voiceflow | 15.8:1 |
| StackAI | 31.5:1 |
| Lindy | 55:1 |
| Gumloop | 87.7:1 |
Botpress has the most balanced crowd in the set, and it is still 11 builders for every service owner or lead. Gumloop's crowd runs 87.7 to 1.
The three at the top of the list sit close together, from 11:1 to 13.6:1. The bottom two are a different category of number. Nothing here inverts.
What the ratio means if you sell into this crowd
If you sell an AI product and your plan is to reach the market through this crowd, the ratio is the ceiling on that plan. On n8n's posts you are in front of roughly 15 builders for every person whose title suggests they sign. On Gumloop's, 87.7.
Builders are a good audience for some jobs. They pick tools and they run the pilots that decide whether a purchase ever gets proposed. In a company of 20 to 200 people they are usually not the person who signs the contract. Two different jobs, and the ratio tells you which one you are talking to.
There is a second reading. When a platform's crowd is almost entirely builders, that platform is being adopted from the bottom up, and content written for a service owner will underperform there regardless of how good it is. Botpress at 11 to 1 leaves more room for a post aimed at a lead.
We are careful with the word buyers, and it does not appear in the table. A title that looks like it signs is not proof that it signs. People whose title suggests they sign is the accurate description, and that is what these ratios count.
The rooms: what this pull did not count
This pull ranked nothing beyond the composition of the crowd. No voices table. No events. No reactions per post.
That is a limitation and printing it is part of the method. For revenue tooling and for ecommerce customer-service AI we ranked the accounts each market follows and looked for the rooms it gathers in. For AI agent platforms at this stage we have who is in the crowd and what their titles say.
What we counted: reactions and comments on eight platforms' public posts in the UK over 90 days, the titles attached to those accounts, and how many people show up around more than one platform.
What we did not count: impressions, follower counts, private profiles, and attendance at anything.
Method: how these ratios were produced
Full method, including where it is weak: /data/method
One question
If the crowd around your category runs 15 builders to every lead, who is your next post written for?
Questions
Because the useful question for this category is who is in the room rather than which vendor holds attention. Reactions per post compares vendors against each other. A ratio describes the composition of the crowd each vendor has gathered, which is what determines whether content aimed at a lead has anywhere to land.
Because a title is a proxy. Somebody whose title suggests they sign may hold no budget for this category, and somebody with a builder title may be the person who decides in a 30 person company. People whose title suggests they sign is the claim the data supports, so that is the claim we print.
No. Bottom-up adoption is how most of these platforms grew, and builders run the pilots that turn into contracts. The ratio matters when you plan content or ads, because a post written for a service owner reaches very few of them on a crowd that runs 87.7 to 1.
Botpress, MindStudio, Relevance, n8n, Voiceflow, StackAI, Lindy and Gumloop. They were chosen as the category's most visible publishers in the UK during the window. Every one of them showed the same skew, which is the finding.
Get this for your market
We build the named market for your category and count who is actually in the crowd, before you spend anything reaching it.