We gave our AI a personality instead of a chat bot window, and it actually changed how people use it.

Building DVOIDER — a platform for turning raw ideas into actual startups. We just shipped something we've been debating internally for a while: our AI companion, Voya.
Most "AI feature" in startup tools ends up being the same thing: a chat window bolted onto the side of the product. Ask it a question, get a wall of text back. We didn't want that, mostly because it makes the AI feel like a tool you consult, not something that's actually part of the product.
So Voya isn't a chatbot. It's shown as an abstract glowing orb, and it does three specific things:

1. Evaluates ideas. When you post a raw idea in D Lab, Voya scores it across 6 dimensions — not a single "good idea / bad idea" verdict, but a breakdown of where it's actually strong and where it's shaky.

2. Tells you what's next. This is the part we cared about most. A score without direction is just judgment. Voya's job is to act like a guide — here's the gap, here's the next concrete step to close it — not just hand you a number and walk away.

3. Matches you with people. Based on the idea and what it's missing, Voya suggests collaborators — cofounders, engineers, whoever the idea actually needs — instead of leaving you to cold-DM strangers.

The "no chat window" decision was the most contested one internally. Some of us wanted a normal chatbot because it's familiar and low-risk. We went with the orb because we wanted evaluation to feel like something happening to your idea, not a conversation you have to initiate and steer yourself.
Still iterating on the scoring dimensions and the matching logic — both are the parts most likely to be wrong right now and we know it.

Curious what people think of AI-as-guide vs. AI-as-chatbot in general — anyone else made this call in their product and regretted (or not regretted) it?