AI in the office comes down to two things in practice: tools that help employees in their daily work, such as summarising meetings and writing drafts, and systems that make the premises easier and cheaper to run, such as booking and energy management. Both can save time, but neither replaces an office that works at a basic level. Here we look at where AI makes a difference today, which rules apply in the workplace and how to get started without buying more than you need.
AI in the office: where does it make a difference today?
Short answer: where work is repetitive, text-based or relies on a lot of data. A simple overview:
| Area | Examples | Who owns the question |
|---|---|---|
| Daily work | Summarising meetings, writing drafts, searching your own documents, translating | You as the employer |
| Administration | Sorting requests, filling in forms, reconciling invoices | You as the employer |
| The premises | Room booking, occupancy forecasts, releasing unused rooms | You, often with the booking system provider |
| The building | Controlling ventilation and heating, early detection of faults in technical systems | The landlord |
For most offices the biggest effect is in the first row. AI in the building is often invisible to the tenant and handled by the property owner.
AI in the workplace: daily work
AI assistants built into office software and meeting tools are already used in many workplaces. The most common uses are to:
- summarise meetings and pull out decisions and actions,
- write first drafts of emails, reports and proposals,
- search and answer questions in the company's own documents,
- translate and simplify text.
What decides whether it works is rarely the technology, but the conditions around it:
- A clear policy on which tools may be used and what information may be entered.
- Order in your documents. AI searching a messy file server gives messy answers.
- Training. People need to know what the tools are good at and where they tend to get things wrong.
- Human review. Anything that goes to customers or authorities, or affects individuals, should be checked by a person.
AI in the premises and the building
In the premises, AI is mainly used to get more out of data that already exists:
- Booking and occupancy. Booking systems can suggest rooms based on meeting size, release rooms that are not used and forecast how many people will come in on different days. That helps you plan and size your next office correctly.
- Energy management. In smart buildings, control systems can learn the building's patterns and adjust ventilation and heating to the weather and expected occupancy.
- Maintenance. Property owners can use sensor data to spot faults in fans, pumps and other technical systems before they cause problems.
Much of this relies on basic sensors and connectivity, which we describe in our article on smart offices and buildings. Without good data, AI features are mostly a label.
What AI does not solve
AI does not turn a poor office into a good one. Noisy open-plan areas, too few meeting rooms, poor ventilation and a location people do not want to travel to are still the most common problems, and they are solved with the right premises and the right layout. Technology can help you see the problems more clearly, for example by showing that meeting rooms are fully booked while open areas are half empty, but the solution itself is physical.
Rules for AI in the workplace
Before introducing AI tools that involve employees, there are a few regulatory frameworks to know about.
The EU AI Act
The EU AI Act applies in stages. Since 2 February 2025 certain uses are banned, including AI systems that infer people's emotions in the workplace (except for medical or safety reasons). At the same time, the requirement took effect that organisations using AI must ensure their staff have sufficient AI literacy. The Swedish Authority for Privacy Protection (IMY) has an overview of the AI Act (in Swedish). AI used to recruit, evaluate or manage employees is classed as high-risk and faces extensive requirements.
Data protection (GDPR)
AI tools that process data about employees or customers are covered by the GDPR. That includes meeting notes and recordings. Biometrics, such as facial recognition in access systems, are sensitive personal data and require very strong justification. An ordinary card or a phone is almost always enough to get into the office.
Work environment and co-determination
New ways of working and new tools affect workload and the work environment. Risks should be assessed as part of systematic work environment management, and if you have a collective agreement, larger changes may also require negotiation with the union under the Swedish Co-determination Act (MBL).
How to get started with AI in the office
- Pick two or three concrete tasks that take a lot of time today, such as meeting notes or proposal material.
- Use tools you already have. Many office suites have built-in AI features. Start there before buying new systems.
- Write a short policy on permitted tools, what information may be used and who reviews the output.
- Train and follow up. Measure whether the task actually gets faster and whether quality holds.
- Expand what works, and stop what does not.
What does AI mean for office space needs?
It is too early to say how much AI will change the need for office space. What is already visible is that the office's role keeps shifting from the place where you sit and produce to the place where you meet, collaborate and learn from each other. That points towards offices with more meeting rooms and social space and fewer fixed desks, a development we describe in our articles on activity-based offices and the future of the office.
Frequently asked questions
How is AI used in the office?
Mainly to summarise meetings, write drafts, search internal documents and handle administration. In the premises, AI is used in booking systems and in the building's control of ventilation and energy.
What does the AI Act ban in the workplace?
Since 2 February 2025 it is banned, among other things, to use AI systems that infer employees' emotions in the workplace, except for medical or safety reasons. AI used to recruit or evaluate employees is classed as high-risk and has specific requirements.
Do we need an AI policy?
In practice, yes. A short policy on which tools may be used, what information may be entered and who reviews the output reduces risk and makes it easier for people to use the tools.
Will AI reduce the need for office space?
It cannot be said with certainty today. What is visible is that the office is increasingly used for meetings and collaboration, which affects how the space should be planned more than how big it should be.
Does your next office need room for new ways of working? Tell us what you need and we will match you with the right office within 24 hours.