With AI nowadays we can create and deploy more product than we can reasonably test and validate. So testing falls to users — and when it does, people often give up on a product instead of coming back to it. I'm flagging this because I think what you've built is good enough to be worth getting right.
Decile group's offering is frankly fantastic — a quick fund setup, a subfund with a licensed GP, and a first-class accelerator program. Product strategy is important, and it should be tied to GTM, branding and operations. That's exactly why the rough edges stand out: there are too many domain names. What's on FI vs decilehub.com vs mybrand.decilehub.com vs Slack — and who to go to for what — is quite confusing.
The standard response to this is more documentation, or a better AI chatbot to route the user. I think that just compounds the problem — you're adding surface to manage surface. The real solution is to refactor: simplify and consolidate the product surface a user needs to interact with.
A couple of directions I'd suggest:
- Treat MCP as a primary surface: Users are more and more working inside their own AI. I pull the sprints down to my own claude. I work on them there, and let it update your database via MCP. My claude will have more context, you will save on tokens, and we follow the old adage "meet people where they are."
- Collapse a surface where you can.
• In web, bring all the hub stuff under one nav, one domain name (decile, not FI).
• This forum already copies the Slack #channel convention — so why not move the whole thing to Slack, or at least mirror it there? That's one less surface for the user maintain, and one less place for users to get lost.
To be clear, this is meant as the opposite of a complaint. The product is strong enough that a little tightening will take it from good to great.