Companion & communication
FAQ & first-contact assistant
Are you open today? Do you have gluten-free dishes? Where can we park? Can we bring the dog? – Every one of these questions is fair. And every one of them interrupts someone in the middle of service.
You'll get …
- 1
Answers in your house style
The assistant knows your opening hours, menu, rooms and policies, and answers guests' questions without interrupting anyone during service.
- 2
From question to booking
Anyone who wants to reserve or is planning a party gets to the reservation or event inquiry in one click.
- 3
A handover, not guesswork
Personal or sensitive questions go to your team – with the request already summarised.
01Intro
The FAQ & first-contact assistant answers the recurring questions directly on your website, in your own style: opening hours including holiday exceptions, the menu with allergens, rooms, cancellation terms, parking. And it takes people further: anyone asking about the function room is pointed to the event inquiry; anyone wanting to reserve ends up in the booking tool. Anything it isn't sure about, it passes to your team – with context, not a dead end.
Worth knowing: the assistant needs a well-kept knowledge base – if opening hours change and nobody updates it, every chat loses credibility. That's why a proper set-up and regular, scheduled updates are part of the project. And with personal or sensitive questions, it hands over to a person rather than guessing.
02How it works
- 01
Collect the questions guests ask most
What do your staff get asked every day, what turns up in emails, what comes in by phone? That gives us the list of questions the assistant has to handle – usually 30 to 50, which cover 90 per cent.
- 02
Build the knowledge base
Opening hours and exceptions, menu and allergens, rooms and capacities, directions, policies: we structure what your place knows – often the first time it's all been in one place.
- 03
Set the tone and limits
Warm or reserved, formal or informal, when should it offer a reservation, and what goes straight to the team? All of this is settled before it says a word.
- 04
Build it in where your guests are
The assistant becomes part of your website – mobile first, because that's where guests check opening hours. In your own design, not as a third-party widget.
- 05
Review and refine
The conversations show what guests ask – which leads to better answers and sometimes a pointer to what's missing from the website. After the first season, the assistant knows your business.
03Stages
The starting point is deliberately limited – in hospitality, an assistant above all has to be dependable, especially on opening hours and prices.
Starter: the FAQ assistant
Curated knowledge on opening hours, menu, rooms and policies, tight guardrails, and a clean handover to phone and email. Reliable rather than all-knowing.
Next step: first contact with handover
The assistant sorts requests and passes them on in a structured form: reservations to the booking tool, events to the inquiry form, special cases to the team – as a usable message, not a chat transcript.
Full build: the assistant that takes action
The assistant deals with whatever comes up in the conversation: it enters the reservation straight into the reservation system, holds the preferred date for a party and lets the team know when an allergy has been mentioned. It also adapts the website to the guest's request, as Neo does on this page.
04Try it yourself
The companion of the fictional “Alte Mühle” event venue. Pick a question – on the right you see where each answer comes from.
A simulation with fixed answers from the knowledge base on the right. With a real language model you can try it with Neo, the companion of this website – bottom right.
The companion invents nothing: it answers from a curated knowledge base built from your content. And for anything that needs judgement or expertise, there's the built-in switch to a human.
The most important part is the least conspicuous: the maintained knowledge base on the left. It decides whether the companion is useful or merely polite.
05Technical implementation
- At its core is a language model with your curated knowledge base as context – at typical volumes it's cached rather than run through a vector database.
- The assistant runs as a component of your website with its own small back end – no third-party branding, no external widget window.
- Connections as needed: a link to the reservation tool, a handover to the event inquiry, a way to contact the team – in the full build, the assistant enters bookings in the reservation system itself, with fixed rules on what it may do without confirmation.
- Guardrails on two levels: behaviour rules in the prompt, plus technical limits in the back end (rate limiting, input limits, moderation).
- Data minimisation by default: only what's agreed gets stored – contact details are kept separately and only with consent.
- Running costs stay manageable: an answer typically costs a few cents – we measure that rather than guess.
for context · This shows how something like this usually gets built – not how it has to be built for you. What actually fits is worked out in the project: around your existing setup, your team, and what stays easy to maintain.