AI workspace

ChatGPT knows the internet. This knows your business.

An AI workspace is an assistant for your team that knows your own numbers, contracts, brand style and agreements instead of the whole internet. You ask a question about your own business and get an answer within seconds, with a source reference. Proven at Inge de Bruijn Zwembad, where the platform makes 100,000+ emails and documents searchable in plain language.

What it does.

01

Looks it up

Finds the answer in your own documents, numbers and agreements, not on the internet.

02

Answers with a source

Gives an answer in plain language within seconds, with a reference to where it comes from.

03

Knows its limits

When in doubt or with sensitive questions, it refers you to a person instead of guessing.

The difference from a free chatbot.

A general chatbot writes nicely, but it knows nothing about your opening hours, your rates or that one contract from 2023. An AI workspace is grounded in your own knowledge. It looks up the answer in your documents and numbers and cites the source, so it doesn't guess. Your team stops searching through folders and inboxes and simply asks the question. Under the hood, that's a language model working on your own knowledge base, which is called retrieval-augmented generation.

What your team does with it.

Anything that now has someone digging through folders, emails or old documents. A selection of what comes up often:

  • Ask any question about your own business and get an answer within seconds
  • Find the right answer in contracts, quotes and old emails, with a source reference
  • Have a business case, quote or report drafted in your own brand style
  • Boil long documents or meeting papers down to the essentials
  • Work in multiple languages without maintaining separate versions
  • Access set per employee, so everyone sees only what fits their role

How does that work, an AI on your own documents?

Your data stays yours.

An AI workspace only works if you're willing to give it your own documents. That's why it's properly set up. Hosting is in the EU, there's a data processing agreement, access is set per employee, and your data isn't used to train AI models. With sensitive or uncertain questions, the workspace refers you to a person instead of guessing. That keeps you in control and prevents made-up answers, which are called hallucinations.

What is an AI hallucination, and how do you prevent it?

How to start, in three steps.

  1. step 1, free

    Free process review

    In half an hour we look at what knowledge is hidden in your business and where your team loses the most time searching. You get on paper what an AI workspace would save you, even if my advice is to hold off for now.

  2. step 2, from €750

    First workspace on your knowledge

    A working assistant on a defined part of your documents, built on your real knowledge. Fixed price up front, no subscription per user per month.

  3. step 3, from €3,500

    Full AI workspace

    All relevant sources opened up, access per role and integration with your systems. After that, maintenance from €99 a month as The Night Shift, cancel any month. The code and documentation are yours.

Fixed project prices. Up front you know what it costs, what I deliver and what maintenance costs after that.

Frequently asked questions

How is this different from ChatGPT?

ChatGPT knows the internet, but not your business. An AI workspace is grounded in your own knowledge base and systems. It knows your numbers, contracts, brand style and agreements, and it doesn't make up answers outside of that. It cites the source with every answer.

What does an AI workspace cost?

A first workspace on a defined part of your documents starts at €750, with a fixed price up front. A full workspace with multiple sources and access per role starts at €3,500, with maintenance from €99 a month. You don't pay per user per month.

Is this safe with business data, and is it GDPR compliant?

Yes, if it's set up properly. I work with a data processing agreement, EU hosting, access per employee and data minimization, and your data isn't used to train AI models. With sensitive questions, the workspace refers you to a person.

Can the AI workspace give wrong answers?

Any language model can hallucinate. A good workspace limits that by drawing only from your verified documents and by referring you on when in doubt instead of guessing. Every answer comes with a source reference, so you can check it.

Who owns the code and the knowledge base?

You do. The workspace runs on your own knowledge base, and the code is handed over in full, with documentation. You're not locked into a platform. You can keep it, change it or move it, even without me.

What knowledge do you have hidden away?

Walk me through one process. In half an hour we look at what knowledge is stuck in your business and what an AI workspace would save you, and you get it on paper.

Walk me through one process