At Pythagoras, we’ve always believed it’s important to own your own property data. With AI, that conviction only grows stronger. Do you already have a clear view of areas, usage and cleaning history? Then you’ve got a solid foundation and an incredible number of possibilities.
This post is about what AI can actually do with the data you already have, and what you could try out as early as tomorrow.
What AI can do with the data you already have
AI for facility cleaning is, in practice, already here. Today it’s used to make cleaners’ planning smarter, direct cleaning based on actual need, streamline scheduling and provide better decision-making support. Without unnecessary technical hassle. It’s about asking the right questions of the data you’re already collecting and letting the answers point to where you should start.
What information is needed for AI to be useful in cleaning planning?
Here are five examples of what you can uncover if you start digging into your own data today. If you can pull these together, you’ll have a really solid foundation for AI to identify efficiency gains.
Unreasonably high/low time spent on a particular area. A room or object may have an allocated time that doesn’t correspond to its area or room type. For example, an office that’s allocated an unreasonable amount of time compared with similar offices at other schools or in other properties. This is often a sign of old schedules that have never been adjusted.
Frequency that doesn’t match usage. Are you cleaning a space five times a week, even though it’s barely used? If you have some proxy for usage, such as booking data or visitor frequency, you can check whether the frequency is actually justified.
Cost that stands out. Cost per square metre or per cleaning visit that deviates from the rest of the portfolio can point to inefficient routes or incorrect staffing levels, rather than to the cleaning time itself.
Travel time that eats into the schedule. Travel time and other overhead between properties often takes up a bigger share of the schedule than you’d think. This pattern is clearest if you have data per shift or round, not just per room.
Quality that isn’t keeping up. If you have any quality follow-up in place, for example under INSTA 800 or your own audits, you can compare it against time spent. This will show you whether there are already rooms where you’re spending less time than required for approved quality, helping you avoid cutting back in the wrong places.
What questions could you ask?
Not sure where to start? Try some of these prompts on your own data. They’re worded to find the same types of anomalies as the examples above:
- Show all rooms where the allocated cleaning time deviates the most from the average for similar room types and areas.
- List spaces that are cleaned more than three times a week but have low booking frequency or visitor data over the past month.
- Calculate the cost per square metre and per cleaning visit for each property, and sort by which ones deviate most from the portfolio average.
- Sum up travel time between properties per shift, and show which rounds have a large share of time going to movement rather than cleaning.
- Compare the time spent per room with the requirement for approved quality under INSTA 800, and flag rooms that fall below the threshold.
Try one prompt at a time and elaborate in the chat if you get follow-up questions. The clearer you are, the better answers you’ll get.
Meet us at Nordic Workplace and talk AI
Would you like to talk more about this? This autumn, we’ll be attending both Nordic Workplace and Lokalvårdsmässan, talking about AI and facility cleaning. Come and meet us there!
