[ad_1]
In this episode, I talk with Mark Sorsa-Leslie, founder and CEO of Auttaa AI, whose team just won first place in Luotea’s inaugural hackathon in Helsinki for turning property maintenance from reactive firefighting into proactive forecasting with AI. Mark is a chartered surveyor with 30 years of real estate data experience, previously the founder of the UK sensor company Beringar.
We discuss why so much building data stays dark and siloed, from outdated networking technology to the shortage of people who understand both engineering and property. Mark explains how Auttaa bridges real-time building data with large language models, using what he calls an “influence graph” to surface correlations and causation across previously disconnected systems, such as linking occupancy patterns to CO2 readings to diagnose ventilation problems in minutes rather than days.
A central theme is Mark’s challenge to the assumption that AI requires perfectly clean data to deliver value. He describes Auttaa’s “messy front end” approach, an agentic pipeline that first characterizes and loads data, then refines it later, flipping the traditional extract-transform-load model to extract-load-transform. We also cover how construction projects could better serve future operations and maintenance needs, particularly by building more comprehensive, asset-complete BIM models at handover.
Mark shares the challenges of being an early-stage startup, from earning customer trust to proving the technology works, as demonstrated during the 48-hour hackathon. He closes with his vision for Auttaa as a software-only, globally applicable layer for understanding buildings and his take on the shifting proptech funding landscape in the Nordics.
Listen to hear how Mark thinks AI can move real estate from teams manually modeling data to trusting large language models, guided by experts, to characterize building systems with a high degree of accuracy.
Connect with Mark at auttaa.ai or by email at .
Listen to the interview and subscribe wherever you get your podcasts
Podcast: Play in new window | Download
Subscribe: