The same details, entered once
An event passes through a dozen stages between the first enquiry and a review. We built a system that carries it through all of them, with people approving what needs judgement.
ResultsThe situation
Every enquiry came in as prose. A date, a rough guest count, a few requirements, sometimes a budget, written the way people write emails rather than the way a pricing sheet wants them. Turning that into a quote meant reading it properly and working through the quoting spreadsheet by hand.
That spreadsheet was not the problem. It held the pricing the business had worked out for itself, and it was right. It just could not do anything on its own. Someone had to sit in front of it for every single enquiry.
And it was one of five. The same handful of facts were retyped into the client breakdown, the contract, the invoice and deposit request, the banqueting and events order, and the pipeline tracker, each at a different stage of the event, each maintained separately, each a chance to get something wrong.
That had a ceiling built into it. Taking on more enquiries meant one of two things: more people to do the work, or the existing team working longer hours to get through it.
What we built
The design question was not whether to automate. It was which steps run on rules and which genuinely need judgement, because those are not the same thing and treating them the same is how these systems go wrong.
Reading a messy enquiry email needs judgement, so that step uses AI, turning the prose into structured detail the rest of the system can work from. Pricing does not need judgement, so it runs as rules taken directly from the quoting spreadsheet, with no model anywhere near the numbers. Nothing was reinterpreted and no rates were invented.
- Enquiries are captured from the inbox or the web form and logged automatically, so nothing depends on someone spotting an email.
- The free text is read into structured detail: event type, date, guest count, location, requirements and budget where given.
- Pricing runs across the full job, from venue hire to supplier costs, with margin checked against target before anything is drafted.
- The quote and its branded client breakdown assemble from that pricing, with AI writing only the narrative sections. Nothing sends until a person has approved it.
- Follow ups go out on day three and day ten, and stop the moment the client replies.
- Acceptance generates the invoice, contract and deposit request from the booking, with nothing re-entered.
- The banqueting and events order fills itself from the same booking, the team is notified, and the pipeline updates itself rather than being updated.
- Afterwards the thank you, survey, review request and testimonial request all go out on their own.
What changed
Ask Rena's what changed and they point to relief. An enquiry is not a booking, so every hour spent pricing one by hand was time given up before knowing if it would pay off. Quotes now go out fast, and nothing about the standard has dropped to get there: each one is still branded and still carries the same detailed breakdown of costs a client would have received before.
Five documents are no longer filled in by hand. The quoting spreadsheet, the branded client breakdown, the contract template, the banqueting and events order, and the pipeline tracker.
The pricing did not change when the spreadsheet stopped being used. That was the point. The rules were taken from it faithfully, so the system prices exactly as Rena's would have.
The hours that went into all of it are spent differently now: on the guest, understanding what they need, refining the details, making the event memorable.