FAQ
Questions and answers.
These are the questions we actually get. If yours is missing, email us and we will answer within one working day.
Getting started
Where do we start if we do not know what can be automated?
That is the normal situation, and it is what the mapping is for. We go through the working day with the people who actually do the work and put numbers on volume and time per step, and you get a list ordered by effect over effort. If you want a feel for it first, the AI Check takes four minutes, see /en/ai-kollen.
What does the first call look like?
30 minutes on video, free, with Caesar. No salesperson and no presentation: you describe where the time goes, we say what can be automated and what it would require. Book it at /en/boka-tekniskt-samtal.
How quickly is something up and running?
A first pilot is normally in production within 2-4 weeks. We start with one bounded workflow that gives a clear effect, then we build further at the pace that suits you.
Who builds the solution?
Caesar and Ted. The same people you talk to in the first call write the code and are responsible for the solution holding up in production. No middlemen and no handovers.
Price and contract
What does it cost?
The pilot is quoted at a fixed price before we start, so you know what you are saying yes to. After the pilot you decide whether we build further, run operations or hand over. The indicative prices are published openly on the service pages, see /en/ai-implementering.
What is included in the fixed price?
Building the bounded workflow we have agreed on, the integrations against your systems, putting it into production and the documentation. What sits outside it is what we cannot foresee, such as a system supplier charging for API access or your data needing cleaning first. If that comes up we say so before we do anything, not afterwards.
What does it cost on an ongoing basis after delivery?
Operations with monitoring, logs and updates is SEK 2,500-5,000 per month, operations with agreed development time SEK 8,000-20,000. On top of that come model calls and hosting, typically SEK 500-5,000 per month depending on volume. The full price list is on /en/ai-implementering.
What happens if the pilot does not deliver?
Then we say so plainly. You have paid a known amount to find out something that was worth knowing, and you decide whether we adjust, switch workflow or stop. We would rather sell an honest no than a project nobody uses.
Who owns the code?
You do. The code, the configuration and the documentation are yours and are handed over in your own repositories. You are not locked into a maintenance agreement to reach what you have paid for.
Technology and systems
Do we have to replace the systems we use today?
No. We build in the systems you already use: ERP, email, documents and internal tools. The automation is connected where the work already happens, so you do not have to learn a new platform.
Which systems can you connect to?
The most common are Fortnox and Visma, Microsoft 365 and Google Workspace, HubSpot, Salesforce and Pipedrive, and Slack and Teams. If the system has an API, or even an export, we can connect it. Unsure about your own system? Ask us, you will get a concrete answer.
Which AI models do you use?
We pick the model to suit the task and switch when something better arrives. Simple classification does not need the most expensive model, while reasoning over a complex set of documents does. We write out which suppliers are part of your particular solution, so that you can take a view on it.
Do we need to understand AI in order to buy?
No. You need to know your own business, the rest is our job. You point at what takes time, we show what can be automated and what it requires.
Security and data
What happens to our data?
Your data stays in your systems. We build with permissions, logging and human review where it is needed, and nothing is moved outside your environment without your approval.
How do you handle GDPR?
We build with data minimisation: only the records the task requires are passed on, and we go through exactly what that means per workflow before it goes live. A data processing agreement is signed where it is needed and we write out which sub-processors are involved. The logging means you can show afterwards what happened.
What does the EU AI Act mean for us?
For most automations of internal administration it means low risk, and therefore mainly requirements on transparency and documentation. We document which risk class your solution falls into and what that means in practice. The point is that the assessment should be made and written down, not assumed.
Can the agent make decisions without anyone looking?
Not by default. Quotes, contracts, customer replies and payments always pass a responsible person before they become binding. Once you have seen enough correct outcomes in the logs yourselves, you can choose to let certain case types through automatically, but that is your decision and it can be reversed.
After delivery
Who maintains the solution?
We do, if you want us to. Maintenance means monitoring, error handling, updates when your systems or the models change, and agreed development time if you want to build further. The indicative prices are on /en/ai-implementering.
Can we take over operations ourselves?
Yes. Everything is documented as we go, not as a closing report, and the code sits in your repositories. Your own IT function or another supplier can take over without us needing to be involved.
What happens when the AI models are updated?
We test against the same set of real cases every time something is swapped out. A new model being better in general does not mean it is better at your particular task. Included in maintenance, see /en/ai-agenter for how we follow up agents in production.