Services03
AI foundation & integration
An assistant that can't reach your documents guesses. We set up a shared AI workspace, connect it to the systems you already work in, and make sure access and logging are in place before anyone types in anything confidential.
AI is only as good as what it can see.
Read onSound familiar?
The answers are generic
The AI tool can't see your own cases, so it answers like a textbook instead of like a colleague.
The material is in five places
Some in the case system, some in email, and some on a shared drive nobody has tidied up in years.
Nobody can say who has access to what
Permissions have grown over time, and nobody dares promise they are still right.
What we help with
A shared AI workspace
Your own workspace in ChatGPT, Claude or another platform, with shared instructions, templates and assistants everyone can use. Instead of every employee having their own account and their own rules.
Your systems and data
Connections to what you already use: case management, email, file archive, CRM. And a review of what is good enough to build on, and what should be left where it is.
Access, logging and traceability
Permissions that follow the ones you already have, and a log you can show. Not a parallel system outside anyone's control.
Running it without an IT department
Set up so an ordinary business can look after it. We don't leave behind anything that needs us to keep it running.
What you get
A shared AI workspace that knows what your business knows, and that you are in control of: who has access, what is logged and who looks after it.
Once the foundation is in place, the answers stop being generalities. That is usually when people stop calling it a toy.
Start here
Internal knowledge assistant
Your shared AI workspace and a first assistant that can find and explain approved knowledge from selected documents and systems. Contained enough to get finished, and real enough to show what the rest will take.
Mapping
We follow a few real workflows from end to end and choose the knowledge the assistant should answer from.
Set-up
Workspace, connections, permissions and logging are put in place in a contained area before anything is rolled out widely.
Testing
The answers are tested on questions you already know the answers to, before anyone uses them for anything that matters. Then one department at a time.
Common questions
Our data is in an old system. Is that a problem?
Rarely one that stops it. The question is usually not whether we can get to it, but whether it is worth moving. We find that out in the mapping.
Will our documents be used to train a model?
Not unless you have decided they should be. The starting point is that your material is used to answer your own questions and nothing else.
We don't have an IT department.
That is the assumption we build for. The solution has to be something the person responsible for IT, who also has other jobs to do, can look after. Otherwise it has been put together wrong.
Let's talk
Tell us briefly about your situation. You'll get an honest assessment of whether there is anything worth pursuing at all.