We use Dashform for project discovery and guided follow-up rather than as another generic form builder.
The feature that made us keep it is the separation between a normal logic-based intake and an AI-assisted follow-up funnel. A fixed form is great when you already know every question you need to ask. In discovery work, however, the useful next question often depends on what the person just told you. Dashform gives us a practical way to test that adaptive layer without replacing the deterministic intake underneath it.
We ran a synthetic discovery proof across multiple scenarios and retained a production Logic Funnel plus a production AI follow-up Funnel afterward. That was important for us: the product had to prove that it could add decision-useful clarification, not simply make a form feel more conversational.
I also like the direction around MCP, webhooks, integrations, analytics, and agent-assisted form creation. Direct create/readback from ChatGPT has been useful, and the product is much more interesting to us because the form can participate in a broader workflow rather than remain an isolated survey.
The main area I want improved is least-privilege integration. During our testing, the MCP/OAuth path exposed rough edges around scope/resource interoperability before the direct ChatGPT path became usable. For an automation-oriented product, the ideal experience would be very explicit read-only vs. write scopes, no unnecessary offline access, and documentation that always matches the live authorization behavior.
I would also love more explainability around adaptive questioning: for each generated follow-up, expose why the system decided the question was worth asking, what missing/ambiguous fact it is trying to resolve, and the stopping reason. That would make the AI funnel easier to audit and safer to use in structured business discovery.
Overall, Dashform has earned a real role in our stack because it solves a different problem from Tally or Google Forms. We still prefer deterministic forms when fixed branching is enough, but when we genuinely need guided clarification, Dashform gives us a useful adaptive layer instead of forcing a human to manually conduct every follow-up conversation.