Ten Questions with Lucy Lyons, CEO at Kestrix

"Imagine seeing a heat map of a home’s energy performance alongside floor plans and photos—it could fundamentally shift how people assess property value and future costs."

What happens when you turn thermography from art to science?

Using drones and AI technology, Kestrix takes daytime and nighttime thermal images of houses, stitches the dual images together, finds where heat is escaping, and then uses its proprietary data to show which houses need which fixes to become more energy efficient.

In essence, Kestrix is creating the world's first AI Thermographer to detect and solve heat loss in housing.

And it's doing it at a potentially extraordinary scale—Kestrix never enters a home, and can scan entire streets and neighbourhoods with a single drone flight. In theory, Kestrix could map an entire city in just one month, and an entire country per quarter.

I spoke to founder and CEO Lucy Lyons (pictured) about topics including AI, mapping properties, retrofitting, portal use cases, and scaling. Here's what she had to say...

What is heat loss, and why is it so important relative to housing?

Heat loss is when purchased energy escapes from a home or building because of poor insulation or inefficient heating systems. It’s a massive part of the climate problem—building inefficiencies are responsible for over 20% of Europe’s emissions. This contributes to higher energy bills as well as health issues for people who live in homes. 

More than 100 million homes in Europe must be retrofitted for energy efficiency before 2050. We already have much of the technology we need to fix the issue (insulation, heat pumps, etc.), but we don’t know which buildings are performing the worst or how to fix them.

We need a trustworthy blueprint to solve this problem.

What problem does Kestrix solve, and in which markets do you operate?

Kestrix solves the problem of not knowing which homes are losing the most heat and what to do about it. We use drones and machine learning to ‘scan’ buildings, creating 3D models that show exactly where and how heat is escaping and generate retrofit recommendations based on that data.

We work with major social landlords and energy companies in the UK who are delivering millions of pounds' worth of retrofit projects. We're also exploring international markets where the retrofit challenges look similar, like northern Europe and parts of the Iberian Peninsula. 

Who are your primary users, and what are their pain points?

Kestrix’s primary users are asset management professionals and retrofit leads at social housing providers, energy companies, and local authorities.

Their biggest challenge is a lack of reliable, property-level data on heat loss—they often don’t know which homes are losing the most heat, or how they’re truly performing. This makes it hard to plan retrofits efficiently or prioritise the right interventions. They’re under pressure to scale quickly to take advantage of subsidies and meet climate targets, but they have limited time and tools.

How is the business funded, and will you fundraise soon?

Kestrix has raised nearly £1 million in equity funding from a mix of experienced angels and the PropTech VC Pi Labs, alongside nearly £1 million in non-dilutive grant funding. This includes support from Innovate UK and the Department for Energy Security and Net Zero (DESNZ).

We’re currently raising to scale our pipeline automation, which will allow us to serve more housing providers and utilities faster at a lower cost.

Our short-term focus is on delivering high-quality data to our current customers and scaling our technical infrastructure. Our long-term vision is a “Google Maps of heat loss”—where every home’s performance can be seen and understood just as easily as an EPC today.

What numbers and findings can you share about your previous and ongoing projects?

We have mapped or scheduled mapping for nearly 10,000 homes on behalf of our existing customers and partners, including EDF, E.ON, EDP, Clarion, Peabody, Islington Council, Lewisham Council and others.

Tell us about your AI. I’m assuming this is a hugely scalable business as long as drones (and their pilots) are available?

AI sits at the core of our business—it’s what allows us to convert 2D images to 3D and pull numbers out of raw image data.

More specifically, we use computer vision and machine learning to align visible and thermal images, generate 3D models of buildings, and estimate heat loss metrics like U-values and energy loss per square metre. What makes our approach special is the combination of aerial thermal data, physics-informed modelling, and automation: we’ve built a pipeline that can process large volumes of data without requiring internal access to the home, and without manual surveys.

And yes, the business is highly scalable—once drone imagery is captured, the rest of the process is automated through our pipeline. That means we can scale across cities and even countries quickly once we find drones and pilots.

Drone technology is core to your business. How quickly could Kestrix map an entire city or country? And what's stopping Matterport from replicating what you do?

Kestrix can map thousands of homes per day using off-the-shelf drones and a trained network of pilots. We’ve already run campaigns that have covered entire neighbourhoods in a matter of days and weeks.

We are building our tech stack to be lightweight, modular, and built to work with a range of drone providers, so we’re able to scale rapidly. In theory, we could map an entire city in under a month and a country like the UK in under a few months with the right partners.

As for companies like Matterport, while they’re great at internal, LiDAR (Light Detection and Ranging)-based scans for commercial or real estate use, they’re not necessarily built for large-scale, external thermal analysis. Kestrix is unique in combining visible and thermal imagery, captured externally only from 50 metres above ground, and processing it through a proprietary pipeline optimised to extract energy performance metrics relevant to retrofit. 

Kestrix works with housing associations today, but surely there is a massive opportunity for portals like Rightmove and Zoopla to buy your data and package it on residential and commercial listings...

Absolutely—and it's a direction we're actively exploring.

While our initial focus has been on solving urgent problems for housing associations and utilities, there’s a huge, longer-term opportunity to surface our data in places like Rightmove or Zoopla because buyers, renters, and lenders are increasingly interested in energy performance, especially with rising bills and regulations on the horizon. As a reminder, EPCs are outdated, inconsistent, and often just wrong.

Kestrix is accurate, visual, and intuitive. Imagine seeing a heat map of a home’s energy performance alongside floor plans and photos—it could fundamentally shift how people assess property value and future costs.

Portals like Rightmove and Zoopla are in a strong position to offer this angle on analysis, and with the right partnerships, our data could be made accessible to millions of consumers in the same way EPCs are today.

Why are you concerned about over-investment in AI agents? Why should investors and VC firms be looking elsewhere?

AI is only as good as the data it learns from. There are real opportunities in agentic AI, especially for automating workflows and improving efficiency across complex systems, but it cannot solve everything.

The problem with housing is that the data is so often outdated, patchy, or just plain wrong. Accurate, primary data at scale is necessary to build game-changing solutions.

This is what Kestrix focuses on: not just using AI, but generating the right data to feed it. Our models are trained on real thermal and visual imagery of actual homes, not guesswork. This is what will enable us to understand the built world properly—AI pipelines that automate the analysis of that messy, fragmented data. 

What isn’t the industry talking about that it should spend more time talking about?

One thing we don’t talk about enough is how we know if retrofit even works for properties. There’s loads of focus on delivery—installing insulation, heat pumps, solar—but not nearly enough on validation. Most of the time, we’re relying on models or assumptions, not real-world evidence. Housing providers are spending millions without clear proof that heat loss has gone down or that homes are genuinely warmer and cheaper to run.

Residents are being asked to go through disruptive works, and landlords are under pressure to hit targets, but without solid data, it’s hard to prioritise, justify, or learn from what’s been done. If we’re serious about scaling retrofit, we need to build in measurement from the start. Otherwise, we risk repeating the same mistakes, just faster.