Sutradhar
An AI copilot for wedding planners that turns WhatsApp chaos into a plan.
A working MVP live for solo freelance wedding planners: structured intake, a completeness copilot, and WhatsApp automation for chasing confirmations, built on a direct Meta Cloud API integration with no middleware.
What it is
The one who holds the thread.
Wedding planners run entire events out of WhatsApp threads and memory. Sutradhar — Sanskrit for the one who holds the thread — structures the intake, flags what a plan is missing, and chases confirmations automatically, so nothing culturally important dies under forty unread messages.
Built with Sandeep Upadhyay and Shagun Jain. I own the product, the AI layer, and the WhatsApp automation design.
How it works
From a rough brief to a chased confirmation
- 1Client brief lands
Rough and messy, exactly as planners actually receive it.
- 2AI structures the plan
Intake turns it into a full event plan. No retyping into a notes app.
- 3Gaps get flagged
A completeness copilot checks the plan against cultural and preference knowledge.
- 4WhatsApp chases for you
Vendor and family confirmations, sent directly through the Meta Cloud API.
- 5The day starts prioritized
Top urgent items across every active event, surfaced each morning.
Product
What it looks like
The pivot
Couples were the wrong user. Planners are.
Built for couples planning their own wedding.
One wedding, ever. No repeat use.
No compounding data, no reason to come back next year.
Built for the solo freelance planner.
Dozens of events a year, repeat use by design.
Every event makes the tool smarter for the next one.
A real market to exist in
Tradeoffs
Decisions I'd defend
Build for the professional planner
A couple is a one-time, low frequency user with no compounding data. A planner runs events repeatedly and the product gets more useful to them the more events it sees. That's a business worth building on.
Integrate the Meta Cloud API directly
Middleware is faster to stand up but adds a cost layer and a dependency I don't control. Direct integration took longer but means the confirmation flows aren't renting someone else's rate limits.
AI flags gaps, a human still confirms them
A planner's judgment on cultural and family specifics is the actual product. The AI's job is to make sure nothing gets missed, not to decide what matters.
Honest gaps
What isn't built yet
AI suggestions still write straight into the plan. An approval queue for the planner is the next fix.
Quality is measured by planner accept, edit, or reject. No labeled precision number yet, and I won't invent one.
Template approvals and pricing sit with Meta. A risk I plan around, not one I control.
“The North Star is events completed on their target date with no missed vendor deadlines — a stricter bar than engagement, and the only number that reflects whether the product did its job.”