AI engineering, measurement and operations
Most AI projects
never reach production.
They rarely die on the model. They die because the solution does not hold up in operation, or because nobody starts using it. We have one person for each.
The whole chain
We work along the whole chain — or on the part where it has stopped.
- Web
- Visibility
- Traffic
- Tracking
- Data
- Conversion
- Automation
- Growth
Three ways an AI project dies — and what stops it
Three areas of expertise under one roof.
Each area matches one of the reasons projects stall. We cover all three.
- 01
AI engineering and data
The solution stays a demo and never becomes a system
Agents, RAG and integrations that reach production — on self-hosted models when the data is sensitive. Plus the measurement underneath, so the effect can be followed up.
- 10+
- years with data and web analyticsSverige
- Matomo · Piwik PRO
- implementation, migration and trainingMyndighet och privat sektor
- Leadsspårning
- through authenticated and identification flowsFintech
We work withAI agents · RAG · Web analytics · Matomo · Piwik PRO · GA4 · Tracking
- 02
Platform and security
The pilot breaks the moment it meets real operations
The platform the AI sits on: cloud, networks, containers and security. Plus the training that lets your own team take over operations afterwards.
- AWS · Azure
- cloud operations certificationsCertifieringar i teamet
- 5,5 / 6
- as a trainerSnittbetyg, certifieringsprogram
- Kubernetes · Terraform
- containers and infrastructure as codeProduktion
We work withAWS · Azure · Kubernetes · Terraform · Networking · Security · Training
- 03
Change management and process
The tool ships, but nobody changes how they work
Getting the organisation to actually move over. Lean and value stream mapping in large, complex environments — the result is measured in usage, not in delivery.
- 61 → 96 %
- system usability after the reworkNordisk servicelinje
- −40 %
- production downtime after integrationGlobal SAP-integration
- 800+
- in the service line managedMedarbetare i linjen
We work withChange management · Lean · Value stream mapping · Project delivery · Scrum
When a single service is not enough
Four situations we recognise.
You rarely need all of it. These are the situations we most often step into — and which roles each one takes.
You want to use AI on material that cannot leave the building.
AI on your own infrastructure
An assistant that knows your internal documentation, with the model running in your own environment. The data never leaves the building — and the platform underneath it is properly deployed and secured.
See scope and setupTakes
- AI and data
- Platform and security
You tried, and it faded out.
AI people actually use
We build the solution and lead the transition at the same time. Follow-up measures usage, not delivery: a tool nobody opens is a failed project no matter how well it works.
See scope and setupTakes
- AI and data
- Change management
You do not want to depend on a consultant.
The capability stays with you
Training and change management in the same engagement, so your team can carry on without us. We have built certification programmes before, not just run courses.
See scope and setupTakes
- Platform and security
- Change management
You have a pilot that has stopped moving.
From pilot to production
The whole way: build, deploy, secure and bring the organisation along. We step in when something exists but has stalled on the way to live operation.
See scope and setupTakes
- AI and data
- Platform and security
- Change management
Sectors we have worked in
The experience comes from somewhere.
- Government agencies
- Fintech
- Listed biotech
- Telecom
- Energy and electricity grids
- Heavy industry
- Global industrial groups
- Grocery retail
- E-commerce
- Education
- Shipbuilding
- Video distribution and CDN
- Management consulting
The assignments were carried out directly or through previous employers. We do not name clients on the website — several of them work in sectors where that would not be appropriate, and references belong in a conversation, not on a public page. Ask us and we will tell you more.
How we work
Fewer assumptions. More evidence.
- 01
Understand the context
We start with your goals, your processes and the data you already have.
- 02
Build what is needed
A scoped solution that works with the tools you already use.
- 03
Measure, learn, improve
We follow up together and refine what produces results.
Start here
What has stalled on your side?
Tell us briefly where things stand and we will come back with a clear next step. No obligations.