Website companion & knowledge chat

01Intro

A companion is an assistant that really knows your business: your services, your processes, your limits. It answers visitor questions in your tone, explains offers, sorts out what someone needs – and leads to the right next step instead of ending at "contact us". A concrete example: Neo, the companion of this website, is exactly that kind of system. He knows every use case here, adapts the page to your concern and remembers what moves visitors. Ask him and you'll know how it would feel for your customers.

Straight up: a companion doesn't give expert advice – no tax, legal or health information; questions like that it hands cleanly to the team. And: it needs a maintained knowledge base (we build that together – that's the real core of the project). Also: what it stores is set beforehand with a data-protection eye, not afterwards.

02Demo

03How it works

1. 01

   Build the knowledge base

   Website, FAQs, services, pricing logic, internal rules: what the companion should know gets gathered and structured. This is where it shows where your knowledge sits – and where the gaps are that visitors feel too.

2. 02

   Define behaviour, tone and limits

   How does the companion speak, what may it do, what never? Formal or casual, when does it offer contact, which topics go straight to a person – that's set before it says a word.

3. 03

   Embed it where visitors are

   The companion becomes part of your website – in your site's design, not as a foreign widget. Mobile and desktop, at the points where questions come up.

4. 04

   Hand over inquiries in a structured way

   When a conversation turns into an inquiry, it reaches you as a usable brief – not a chat transcript. Unclear and sensitive cases go straight to a person.

5. 05

   Watch and sharpen

   What do visitors really ask? The conversations show it – and out of that come better answers, new FAQ entries and sometimes a new offer. The companion gets more useful every month.

04Stages

The entry point is deliberately limited – a companion doesn't have to do everything on day one, it has to be reliable.

05Technical implementation

• At its core is a language model with your curated knowledge base as context – at typical volumes cached rather than run through a vector database (more honest and cheaper).
• The companion runs as a component in your website with its own small back end; no foreign branding, no external widget window.
• Tool connections as needed: inquiry handover, appointment hints, page recommendations – on this site also state changes and signal capture.
• Guardrails on two levels: behaviour rules in the prompt plus technical limits in the back end (rate limiting, input limits, moderation).
• Data minimisation as the default: what's stored is what's agreed – contact data kept separate and only with consent.
• Running costs stay manageable: a warm answer typically costs a few cents to run – that's measured, not guessed.

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.

06Related building blocks

• Knowledge

  Knowledge catalog

  The foundation under the companion: services, prices, FAQs and processes as an ordered knowledge base.

• Communication

  Inquiry router / offer finder

  The little sibling: when visitors mainly need the right path, a guided router is often enough – no free conversation at all.

• Website

  AI-ready website upgrade

  The step before: structured website content is exactly the knowledge the companion later answers from.

• Sales

  Inquiry qualifier

  The counterpart after the conversation: the visitor's concern becomes a structured brief for your team.