Dodoform
AI forms that turn voice, photos & notes into clean data
Start new thread

DodoForm v2 - AI forms that turn voice, photos & notes into clean data

by•
DodoForm 2.0 rebuilds intake around AI. New since v1: live voice-to-text that types straight into the form, photo/file extraction, a human review queue so AI never silently guesses, conversational analytics ("ask your data anything"), a dashboard Extractor for pasting messy data, and an AI-credit model with unlimited responses.

Add a comment

Replies

Best
Maker
📌
I'm building Dodoform because every form I'd ever used fought the way people actually communicate. People don't think in tidy fields — they leave a voicemail, forward an email, or snap a photo of a form. Someone on the team then re-types all of it. Dodoform flips that. Respondents can type, speak, snap a photo, upload a file, or just paste messy text — and AI extracts it into the right fields. You get a review step to verify anything low-confidence (no silent hallucinations), then clean records route to your CRM/ATS/webhooks. A few things I care about and built in: -Unlimited responses on every plan — you're metered on AI work (credits), never on how many people reply. -If you run out of AI credits or a model hiccups, the response is still captured — never dropped. -200+ templates, 25+ field types, conditional logic, multi-page forms. -Conversational BI — ask "what do enterprise users complain about most?" in plain English. Free plan is genuinely usable (3 forms, 100 AI credits, webhooks, integrations, CSV — no card, and yes, unlimited responses). Would love your honest feedback, especially on the voice + photo intake. I'm here all day to answer anything.

How does the AI handle ambiguous inputs like a voice note where someone rambles before giving their actual name and contact info — do you end up needing to verify most fields anyway, or does it get enough context to fill them confidently?

Maker

 It usually gets enough context to fill obvious fields confidently -name, email, phone, company, intent, etc. even if the person rambles first.

But Dodoform doesn’t blindly trust ambiguous extraction. It separates fields by confidence: clear fields are filled automatically, uncertain ones are flagged for review. So you’re not verifying everything ,only the parts the AI is unsure about.

How does the AI actually handle edge cases where the uploaded doc is genuinely ambiguous, like an invoice with a missing field — does it flag it for review or just guess?

Finally tried this and the voice-to-field thing actually worked on my mumbled test answer, which I did not expect. Liked that it shows the raw input next to the cleaned fields so I can sanity check before it pipes to my CRM.

the "verify in one click" step before sending to the CRM is such a smart guardrail. feels like you actually thought about the trust problem most AI form tools hand-wave past.