What does your company do?
We measure architectural quality. We take all the information available and turn this into spreadsheets to give our customers a better insight into their data, so that they can make better business decisions.

Do you have a background in architecture?
Yes, I have my diploma in architecture. Before that, I did a degree in mechanical engineering. In the end, I did my PhD at ETH in Zurich in machine learning for architecture.

So Archilyse combines these different areas?
Exactly. We bring tools from mechanical engineering into the architecture domain. All those tools are based on machine learning, so this is the fusion of the three fields I studied.

Why is your solution a novelty?
There was no way of measuring architecture quality objectively. When we talk about architecture, we always talk about “gut feeling” and experience; there are no hard metrics. With our bullet-proof, hard metrics we can now make proper decisions in the architecture domain. For example, we can tell our customers whether changing the floor plan slightly will affect the rental price.

How is this a step up from the existing solutions on the market?
Before we came along, a realtor would visit a unit with a checklist with three levels for floor-plan or view quality and say whether it is mediocre, average or better-than-average. Assessments differed a lot. We now use mathematics to do a proper analysis of the whole context.

How many parameters does Archilyse consider?
Our most advanced customers use about 150-250 parameters actively, but we can provide 15,000.

What are the most important ones?
Sun and view. We can tell you exactly which mountain peaks you will see from your couch, or how much of your field of vision will be occupied by a lake view. Those parameters are not just important for pricing; they affect how people feel. We can measure sunlight, too. Bright rooms are important for children’s cognitive capacity. Apartments with good evening sunlight in winter do better on the market than those without it. This is something that no realtor could take into consideration, because it is hard to tell whether the sun shines in the apartment in the evening in winter.

Isn’t the view often beyond people’s control?
Sometimes. Large developers mostly build four or five buildings on a site, which gives them more degrees of freedom in how to optimise it. They can also use our tool earlier, when looking for a site.

What parameters inside the apartment do you consider?
Different view conditions and logistics; where you can walk, the distances and the adjacencies. Is there an outdoor space next to your dining space, or do you have acoustic separation between your living room and your children’s bedroom? How well is it suited for multiple furniture layouts? Other parameters include wheelchair accessibility and compliance checks.

Why was it so difficult to account for these variables in the past?
It is a difficult mathematical problem, especially with our holistic approach; rather than just one single element, we do them all at once. There were some ideas in academia, but nobody was able to close the gap between those features and economic KPIs like rental price. By closing this gap, we generated a much broader market for those insights.

It sounds technologically complex.
We have different machine learning setups. I wrote my PhD thesis on this topic: architectural descriptors.

Why is this “holistic” approach, as you describe it, so valuable?
This is how we perceive space. I assume that you are sitting in an office and can look out of a window. The view outside somehow belongs to the space in your office. If you look out onto a brick wall or a lake, you perceive your office differently. We can measure this scientifically. If you want to know how a building performs and how people behave and feel inside it, you need to cover both internal and external parameters.

Who are Archilyse’s target customers?
Valuers are the most obvious group, as we replace the manual process of objectifying architectural qualities. Asset managers use our software to better set rental prices. By setting them more in line with market expectations, you can increase revenue by 2-4%. Developers and planners need this information to optimise and conduct the initial rental price assessment. Finally, large corporate real estate owners that want to optimise their own portfolio.

How does your software help companies make informed decisions?
We give them the raw data plus dashboards, with nicely clustered aggregates. At the atomic level of planning, we can give insights into what you would need to adjust to have a better-performing apartment, which means that there is a higher need for this kind of apartment on the market. We close the gap between tenants and planners.

Does your tool have the potential to change how developers think?
It is already. We are turning architecture, especially for “anonymous buildings” like offices and housing projects, into an engineering science. This leads to a totally different approach. We are the key enabling technology for this kind of decision-making.

How does this change the financial calculation for companies?
We can turn floorplans into financial spreadsheets, and vice versa. We can tell them what to change in the floor plan to have better numbers.

Does this reduce the financial risk for them?
By far. We do not have accurate numbers yet, but we see that it is reducing the risk.

What are the implications for companies’ profits?
We are generating added value, adding a profit of several percent, while decreasing costs by a factor of ten. Some of the assessment work used to be done manually; now these parts can be automatized. Companies have higher revenue and lower costs, so they have much higher profits when they work with us, throughout the value chain.

Archilyse is already working with leading real estate companies like Swiss construction giant Implenia. How are you helping it plan ahead?
We are analysing every construction site. We have automatized existing processes and are providing much better planning insights about how the units will compete on the market. Developers give us the floorplans of a new building site and we generate insights in how much every unit is worth, as well as which features can increase a unit’s value. For example, in this unit it would help to make the balcony larger or to add another bathroom.

How is your tool being used by Swiss Life, the life insurance company?
We are digitising the Swiss Life portfolio of 35,000 units. With these insights, the company thinks that it can increase rental revenue by 2-4%. This gives a return investment of a few thousand percent.

How is it helping Credit Suisse invest in real estate?
Credit Suisse, which we work with, develops for sometimes CHF 1 billion a year. Every new site is a high-risk investment for it at the beginning because it does not know what the rental price will be. We have shown that the models based on our features perform much better than everything else on the market.

How will your solution be used by ImmoScout24, Switzerland’s most-visited online marketplace for residential and commercial property?
ImmoScout is enriching its listings with additional information, some of which will come from our API. This new feature will go live very soon.

What about applications in the public sector?
We work with the Bundesamt für Wohnungswesen, the Swiss housing authority. When you want to get additional funding from the Swiss state for an amazing floorplan, they test it. We currently analyse how more of these tests could be done with our APIs.

Do you have plans to expand beyond Switzerland?
We are already working with customers from Germany and France, and have some leads from the UK, the Nordics and Austria. We want to be active in at least four countries next year.

What plans do you have for 2020?
Scaling. We are increasing our speed rapidly and, for me as the founder, this is the most exciting thing.

Archilyse is a portfolio company of Swiss Immo Lab. 

Verve Ventures acts as the investment manager for Swiss Immo Lab.