An AI-powered venue finder that answers the questions brochures leave out
When my partner and I started looking for a wedding venue, the brochures were full of photos and short on the answers we needed. Could we hold a havan indoors? Could we bring our own caterer? What would corkage cost, and when would the music have to stop? Each answer took an email or a phone call, and plenty of them ruled a venue out. So I built Asian Wedding Society, which answers those questions up front for 125 venues and matches couples to the ones that fit. I designed and built it, and I run it.
- Company
- Asian Wedding Society
- Year
- 2026
- Role
- Founder: product, brand, design and engineering
- Team
- Just me, with Claude agents
- 0
venues, with fees, timings and rules each checked against a source
- 0
people visited in the first three weeks after launch
- 0
people planned their wedding with the AI chat, sending 535 messages
The problem
The details that rule a venue in or out for a South Asian wedding (havan permission, dry hire, caterer rules, bar fees, the music cut-off) are buried in PDFs and email threads. Couples shortlist on photos, then spend weeks chasing answers that rule half the list out.
My aim was for a couple to tell within ten seconds whether a venue is worth pursuing. Where we don't have an answer yet, the page says ‘To confirm’ instead of guessing.
Starting with a conversation
Couples start by answering questions in a chat, one at a time, with suggested answers they can tap. Their answers build a wedding brief covering guest numbers, budget, location, ceremonies and catering. The chat also works by voice.
Couples can edit the brief and share it with their partner, and every venue is ranked against it.
Keeping 125 venues accurate on my own
Claude agents find new venues, read their brochures, and check and repair records. Four times a day they email venues with any outstanding questions, apply the replies to the data with the source recorded, and send me anything they're unsure about.
Removing a feature that didn't work
My first instinct was that couples would want to describe their wedding in their own words and let AI find the venues, so I built that. In use, results took up to 12 seconds and the matches weren't relevant enough. Plain, fast filters did the job better, so I took the AI search out.
Early traction
The site launched in mid-October, and two thirds of visitors arrive on mobile. Each week I read the analytics alongside the planner conversations to decide what to work on next.
- Visited · 757 sessions565
- Viewed a venue · 885 venue pages230
- Used the AI planner · 535 messages42
- Created an account10
- Sent a venue enquiry6
Instagram views in a single week, with 175 followers
venues answering the questions our agents send by email