Event requests: reply faster, lose fewer deals
A hotel group receives over 3,000 seminar and event requests every year, handled differently from one property to the next. We structured how these requests are received, priced and followed up, to free up sales time and win back business that was being lost.
The company
A hotel and events group with 4 properties, receiving 3,200 seminar and event requests per year.
The problem and its cost
Requests arrive by email, web form, phone and booking platforms. Each property logs them in its own way, checks availability by hand and builds every offer from a blank document. A request takes 25 minutes on average, often because information has to be asked for again. Follow-up after the offer is sent depends on individual memory.
This has a cost that can be quantified. Handling time represents about 93,000 CHF per year (3,200 × 25 min ÷ 60 × 70 CHF). On top of that, some business is lost for lack of a fast reply or a reminder: we assumed 2% of requests, or 64 events, valued at 6,500 CHF with a 35% margin, about 146,000 CHF. The annual cost reaches 239,000 CHF: 3,200 × 25 min ÷ 60 × 70 CHF + 3,200 × 2% × 6,500 CHF × 35%.
The strategy
We set the rules first, before touching any tool. We had to define what a complete request is, which questions to ask, who replies and within what time. Without this shared frame, automation would only have reproduced the differences between properties.
We then standardised pricing, put follow-up in place, and only at the end added AI to reading and drafting tasks. This order lets each step be measured: first-reply time, conversion rate, reasons for loss. It also keeps the sales team in control of anything that commits the group.
The execution
Team: A dedicated contact, the WALVIS core team (strategy, finance, operations, data, AI) and a sector expert profile in event sales for multi-property hotel groups.
- →A single qualification grid for requests, with the information to collect at first contact and rules for allocating requests between properties.
- →A library of offer templates by event type, with pricing rules and terms approved by management.
- →A shared tracking board built in the group's existing tools, with scheduled reminders, response time and reason for loss.
- →An assistant that reads incoming emails, extracts key details and prepares a draft first reply, reviewed by the salesperson.
AI reads requests received by email, extracts dates, headcount and needs, then prepares a draft reply. It sends nothing on its own: every offer is validated by a member of the sales team.
Most of the gain comes from clear rules, offer templates and systematic follow-up, which need no particular technology. We worked with the tools already in place and added no software.
The result
Expected annual saving: 50% of CHF 239,000, about CHF 119,000. For an investment of about CHF 32,000, payback comes in 3 months.
We count on 50% of the cost rather than 100%, because some requests will always need a human conversation and a tailored quote, and because not every lost deal is lost to delay or poor follow-up (price, unavailable dates). The expected saving is about 119,000 CHF per year for an investment of about 32,000 CHF, a payback of roughly 3 months.
- A consistent first-reply time across the four properties, which clients read as a sign of reliability.
- Sales teams who spend more time on complex events and on client relationships.
- A clear view of the pipeline and of reasons for loss, useful for adjusting the offer and pricing.
The assumptions
- Volume: 3,200 requests per year, 25 minutes of handling per request.
- Loaded hourly cost: 70 CHF.
- Average order value: 6,500 CHF, with a 35% margin.
- Business lost for lack of responsiveness or follow-up: 2% of requests.
Scenario type, built from real missions and anonymized.
In the same cocoon
