Event configurator built for a hotel group
Case study · Hospitality & MICE

Event Configurator for a Hotel Group

A seven-step journey on the hotel group’s own website: it returns a proposal to the client and an already-qualified request to the sales team. From an email with a PDF to a qualified lead.

Summary

A Portuguese hotel group with 8 properties across 3 regions replaced its email quote request with an event configurator hosted on its own website. The client moves through seven steps — region, event type, details, room, catering, proposal and booking — and receives a proposal document in six languages at the end.

The sales team no longer receives "I'd like to know prices for an event". It receives a lead with the room, the date, the headcount and the catering choices already recorded. The catalogue covers 135 event spaces and 269 catering items, managed through six internal back offices with region-scoped permissions.

Figures as of September 2026. The client is a Portuguese hotel group, not named by editorial decision.

The problem was not the price, it was the question

In a hotel group, an event enquiry almost always arrives the same way: a short email, no firm date, no settled headcount, and no idea which space. From there the sales team does all the work — calls, asks, proposes rooms, builds a quote by hand, sends a PDF, waits.

The cost is not the time spent building the quote. It is that the same work is repeated for enquiries that were never going to convert, and that nobody knows how many clients dropped out of a conversation that never got started.

The configurator moves qualification to before the first human contact — without turning it into a form. The client answers the questions because they are the journey, not an obstacle placed in front of it.

The client journey, in seven steps

Each step only asks what the previous one made it possible to answer.

  1. Region — Algarve, Lisbon or Cascais-Oeiras. Up to three at once.
  2. Event type — corporate, wedding, social or other, each with its own sub-types.
  3. Details — date, headcount, room layout, accommodation, equipment.
  4. Room — the spaces that fit, ranked, with photos, capacities and rates.
  5. Catering — packages, meals, drinks and decoration for that room and that hotel.
  6. Proposal — the summary with the estimate, before asking for contact details.
  7. Booking — billing, confirmation, and the generated document.

A room's capacity is not one number

It is eight, one per layout — theatre, classroom, U-shape, cocktail, banquet, cabaret, buffet and single table — and it is the requested layout that decides whether the room fits.

A room seating 400 in theatre style may take 235 in classroom. A request for 300 people for a training session rules it out, even though the brochure says 400. This is the kind of rule an experienced salesperson carries in their head and a generic form ignores — and it is why mapping the process comes before any technical decision.

The inside: six back offices

The configurator is half the work. The other half is six back offices, each with its own audience: whoever manages the room and F&B catalogue, whoever manages the application experience, whoever writes the document copy, whoever tunes the recommendation rules, and whoever works the incoming leads.

Permissions are not only by role — they are also by region. An Algarve manager sees and saves the Algarve; what they edit does not touch what someone else is editing in Lisbon. In a group with properties spread across the country, this stops being a convenience and becomes the condition for several people to work on one catalogue at once.

The recommendation engine is deterministic, and that is the feature

The rooms proposed to the client do not come from a model. They come from a weighted score across factors — fit to the event type, capacity match, region, equipment — whose weights sit in a panel the team tunes on its own, with a test mode that shows, room by room, why each one was included or left out.

A hotel group cannot tell a client that a room did not appear because the model decided so. Here, they can open the panel and show the line that excluded it. It was a deliberate choice: where explainability is worth more than sophistication, the rule beats the model.

One pricing engine, not five

The same figure appears in the step 4 estimate, the step 5 catering, the step 6 summary, the step 7 booking and the final document. Five screens, one source of calculation — shared, including with the internal tools that regenerate the document for an old lead. When a rule changes, it changes everywhere or nowhere.

The scale

8
Hotel properties
135
Event spaces in the catalogue
269
Catering items
6
Languages served
6
Internal back offices
7
Steps in the journey

Figures as of September 2026, verified against the catalogue on the date. The catalogue is live: of the 135 spaces, 65 are active, and catering items grow with each new season.

Frequently asked questions

Is an event configurator worth having on a hotel website?

It is worth it when the volume of enquiries justifies the manual qualification work. The gain is not selling more online — it is the sales team receiving requests that are already answered, and no longer rebuilding quotes for enquiries that were never going to convert.

Does the client book the event alone, without speaking to anyone?

No, and that is not the aim. The journey ends in a proposal and a recorded request; confirming availability and closing remain with the sales team. The configurator shortens the conversation, it does not replace it.

Do you need artificial intelligence to recommend event spaces?

No. Here the recommendation is a weighted score across factors, chosen deliberately over a model: it is explainable line by line, the team tunes it without a developer, and the same request always returns the same result.

How do several properties manage one catalogue without overwriting each other?

Through region-scoped permissions. Each manager sees and saves only their own region, and a save never overwrites what someone is editing elsewhere — a practical condition for a group with properties in different parts of the country.

Which languages does the configurator support?

Six: Portuguese, English, Spanish, French, German and Mandarin. Not every field exists in all six — what the client sees along the main journey is translated, and the detail content stays in Portuguese and English, by editorial decision rather than technical limitation.

Who did this work

BigLearn is a Portuguese artificial intelligence consultancy, founded in 2017 and based in Lisbon. We work with organisations that already have working processes and want to know where AI improves them — hospitality included.

This case came out of our AI consulting for companies and AI agents and business automation work. Every project starts with a proof of concept with a written success criterion agreed before we begin — and if it does not pass, we say so.

We work in hospitality and tourism, local government and public administration, insurance, healthcare, manufacturing and professional services. The other case studies are published under the same rule: client anonymised, verifiable figures, and the nature of the document stated up front.

Do you have a process that looks like this one?

Tell us what happens today and what it costs you. We will tell you whether there is a case worth building — and if there isn't, we will say that too.

Present your case