About AI.
How AI works, what the rules say and how to get started. Explained simply.
The basics
What AI is and when its answers can be trusted.
- What is a language model? A plain explanation What a language model is, what it does well, where it falls short and what that means at work. Explained without technical detours.
- What is an AI agent? A simple explanation What an AI agent is, how it works and how it differs from an AI assistant. On the language model, the tools, the memory and the loop that holds them together.
- How to stop AI from making up answers Why AI sometimes makes things up, and how to reduce the risk with your own documents, sources and clear control.
- Why sources make AI answers trustworthy An AI answer without a source has to be taken on trust. How source references make AI answers checkable, and why that matters at work.
- Train AI on your documents or let it read them? Should AI be trained on your documents or read them when a question is asked? RAG and fine-tuning, and what the choice means for currency, sources and access.
- The difference between internal AI and public AI services Internal AI or public AI services? When the work involves your organisation's own data, the differences become clear: data, training, permissions and sources.
Security and privacy
Where your data lives and who can reach it.
- Is your own installation or a cloud service the better fit? Should your AI assistant run in your own environment or be run by the supplier? How to choose between your own installation and supplier-run hosting.
- AI for organisations with strict confidentiality needs Public authorities, law firms and insurers can use AI without giving up control. Own installation, air-gap, access control, traceability and sources.
- Permissions in AI decide who sees what An AI assistant must never become a shortcut around your permissions. Roles, sign-in, sharing and traceability, and what to require.
- The risks of AI agents and how to manage them What are the risks of AI agents and how do you manage them? On security around hidden instructions, permissions and actions that cannot be undone.
- How to use open-source AI safely in your organisation Open-source AI gives you transparency, but responsibility for what you download is yours. On add-ons, packages, updates, model files and supply chain rules.
- How to handle AI agents that contact you Emails, forms and calls may come from an AI agent. On knowing who is behind it, protecting your systems and handling orders, data requests and fraud.
- When staff type sensitive information into AI What happens when staff paste customer data into public AI services? The risk, why bans rarely work, and how to make the secure choice the easy one.
- What happens to our data when the contract ends? The question often forgotten when choosing an AI solution. Export, deletion, your own installation and what to require before you sign.
Law and regulation
GDPR, the EU AI Act, NIS2 and what US law means for data in Europe.
- US law and data stored in Europe What the CLOUD Act, FISA 702 and the EU-US framework mean for data stored in Europe, and how to decide which data belongs where.
- What the CLOUD Act means for your data US providers may have to disclose data stored in Europe. What the CLOUD Act says, how it meets the GDPR and what to ask your supplier.
- FISA 702 and what intelligence laws mean for the cloud FISA 702 and EO 12333 govern US intelligence gathering outside the United States. The position in autumn 2026, with dates, and what it means for cloud services.
- Is the EU-US Data Privacy Framework enough? The framework eases transfers of personal data to the United States and applies today. It is now before the EU Court of Justice, so a plan B is sensible.
- Cloud services and confidentiality in the public sector What the Swedish secrecy rules, eSam's guidance and the government's cloud policy mean when authorities, municipalities and regions choose cloud and AI.
- When is an AI answer an official document? What Sweden's public access principle means when authorities use AI. Prompts, answers and logs as official documents, confidentiality and the ongoing inquiry.
- Is a sovereign cloud enough? US providers offer European clouds with operations in the EU. They settle a great deal, and one legal question has yet to be tested in court.
- GDPR and AI: the questions to ask your supplier Using AI with personal data? The GDPR questions to ask your supplier before you choose one.
- What the EU AI Act means for your AI assistant What does the EU AI Act require of an organisation using an internal AI assistant? Risk levels, transparency, AI literacy and the dates that apply.
- What NIS2 and the Cybersecurity Act mean for your AI tools Sweden's Cybersecurity Act now applies. How NIS2 affects your choice of AI tools: supply chain, access, logging and incidents.
- What DORA means for AI in financial services What DORA requires when financial entities use AI from an outside supplier. The register of information, exit strategies, concentration risk and supervision.
Sector by sector
How AI is used where the demands are high.
- AI for primary care gives more time for the patient How a primary care centre can use AI day to day. Draft notes and referrals from dictation, answers from your own routines and support for all staff.
- What does secure AI in healthcare take? What does it take to use AI safely in healthcare? Patient data, responsibility, access and traceability, and what to ask a supplier.
- AI for law firms with strict confidentiality needs AI for law firms with client files kept in your own environment. Permissions per matter, answers from your templates and memos, and the lawyer in charge.
- Confidential, accountable AI in the public sector How the public sector can use AI without data leaving the organisation. Confidentiality, accountability, own installation, access, traceability and sources.
- Secure AI for the financial sector How the financial sector can use AI without sensitive data leaving the organisation. Own installation, air-gap, access control, traceability and sources.
- AI for HR and staff questions How HR can use AI to answer staff questions faster, without spreading sensitive data. Material, permissions, boundaries and getting started.
Getting started
From the first decision to part of everyday work.
- How to roll out AI search in your organisation, step by step A plan for rolling out AI search across your own documents, from the first use case to the whole organisation using it.
- Checklist for procuring AI What to ask before you choose an AI solution. A checklist on data, security, access, traceability, answer quality and exit.
- How to write an AI policy your staff can follow How to write an AI policy that works day to day. Approved tools, which information may be used, responsibility, AI literacy and shadow AI.
- The cost of not using AI The hidden cost in many organisations is the time it takes to find the right information. How to make it visible, and how AI can give that time back.
- Onboarding new employees with answers from day one How new employees find procedures and answers on their own from day one, with sources and the right permissions, without constantly asking colleagues.
- How to keep the knowledge when experienced staff leave When experienced staff leave, knowledge that was never written down often goes with them. How to make it available to everyone, from your own documents.
- Tenders and proposals with AI in your own style How to write tenders and proposals faster with AI: find earlier answers and wording, write in your own style and keep the material confidential.
Want to see how it would work for you?
We’ll show you iSearch with your own documents.