Name: Andrew Thompson
Job title: CTO at Orbital
What I do: I am the chief technology officer of Orbital, a legal AI platform built for the real estate industry, automating the legal work behind property transactions, used on over 250,000 deals a year across the US and UK.
Why I got into the property industry: My background is in engineering and I’d spent years in software and biomedical before I went anywhere near real estate. I first worked in property in Vancouver, working on a social platform for real estate agents. Up to then, I hadn’t been involved directly in the business of real estate, but that changed when I stepped into my role as CTO at Appear Here, a proptech company which was essentially an Airbnb for the high street.
While there, I worked with landlords in London, Paris and New York, helping brands launch pop-ups and short-term retail. It was a fast-growing business and running it meant that I was dealing with real estate transactions every day. Then when Covid hit and the high street shut overnight, our whole model got tested very quickly, and we had to cut deep to survive. Around the same time I was introduced to Will Pearce, Orbital’s CEO. At Appear Here I’d dealt with real estate transactions, so I knew the problems he was dealing with first-hand.
Real estate is the world’s largest asset class, yet the legal workflows around transactions are mostly manual and require document-heavy review across hundreds of interdependent records, maps and historic deeds (many of which date back to the 19th century). The industry needed to be changed and from an engineering perspective it was a hard machine learning problem that I was excited to solve. After a few conversations with Will and the team, I joined as CTO.
How I did it: I’ve always been interested in hard yet interesting problems. I’d been a founder before and by the time I was CTO at Appear Here I’d worked across engineering, product and the commercial side of property deals. That combination of both understanding the technology and the day-to-day reality of real estate transactions was fairly unique and it put me in the right place just when legaltech took off as a sector.
My first proper job: My first official job post university was actually in Edinburgh for a company initially called Barco Advanced Visualization which weeks after joining got acquired by Toshiba and became Toshiba Medical Visualisation Systems Europe (TMVSE). I was a software engineer there on a team building radiology software to visualise images that come out of MRI and CT scanners so that radiologists inside hospitals could diagnose patients before going in for operations.
My first big break: Emotionally for me this was when a British chap called Tom Rossiter, who was living in Vancouver, got in touch with me over Linkedin and asked me to go for a pint and a beer to chat about a little startup he was CTO of and wanted me to join his team at RESAAS. I joined them and helped him grow the team. That quickly turned into me becoming VP of engineering and VP of growth. I got thrown in the deep trying to figure out how to grow a startup from nothing.
I learnt a ton, made a ton of mistakes but Tom and team gave me the room to grow my first big role where I was directly responsible for people, technology, product, marketing and overall innovation. I then immigrated to London with my wife and nine month old son. It was a dream of ours to eventually settle in London after previously living in Edinburgh for a year. I arrived in London still quite young and needed to establish myself in a brand new city, but I had the confidence off the back of being VP of engineering and growth and used that to work at a variety of startu-ps in the city and eventually start my own AI one.
My inspiration(s): I find working with AI incredibly intellectually rewarding and I’m an engineer so the ability to apply AI to real-world problems motivates me every day. The idea that I can build something today that might still be operating a decade from now impacting the world is pretty incredible. My blood, sweat and tears today gets embedded into a product that outlives my own effort.
The biggest challenge I’ve encountered to date: Orbital originally tackled legal reasoning with classical machine learning models. At an early point in our product development, I had a team of 10 paralegals labelling clauses in legal documents to train our models. When large language models (LLMs) arrived, we made two bets: that token costs would keep falling until the product was commercially viable, and that AI model capability would keep improving until complex real estate legal reasoning was within reach. Neither was a sure bet.
Sunk cost is not a strategy, so we scrapped a huge amount of that earlier machine learning work and rebuilt everything with an AI-first mentality which put LLMs and, later, agents at the core of our product strategy. At the time it was very much a leap of faith given that OpenAI with GPT-3.5 was the only offering available. Convincing a company to jump with you is harder than any technical problem, but it was well worth it.
My greatest achievement/proudest moment to date: Taking a huge risk on LLMs and agents so early on before it was clear how important they were going to be for application development. It eventually paid off and with hindsight it seems obvious, but at the time we had to take some big risks that could have been very costly and had to figure things out from scratch before the world had woken up to this whole era of AI development.
The biggest misconception people have about the property industry: People working in the real estate legal industry already deeply understand this, but for people on the outside it’s not always clear how important the legal framework is that underpins what can and cannot be done with a property. This framework supports or hinders how properties, neighbours, cities and the countryside are used, redeveloped, revitalised, etc.
It’s sometimes said that “legal is the foundation of capitalism” and real estate legal is the cornerstone of how our cities keep their history, modernise where necessary and ultimately allows us to adapt the built environment around us to the values and culture we adhere to and create over time. The home we live in, the coffee we buy from our local cafe and the office we go to work in would not exist in their current form without the guarantees discovered and afforded throughout the legal real estate process.
Where I’d like to be in five years’ time career-wise: Five years from now is a very long time in this new AI-era and I expect the world to look a bit better than it does now work wise but succinctly I’m an engineer at heart. That means I want to be solving problems in novel ways using the latest technology whatever that latest technology happens to be five years from now.
Who I’d invite, living or dead, to a dinner party: My friends and family!
What I do in my downtime: Novelty is the spice of life so I’m always up to something new, but the fairly consistent ones are spending time with my family (wife and two children), cycling the Surrey Hills or Yorkshire Dales with my son or hitting up my local Curzon to watch the latest independent film.
The job I’d be doing if I was not doing this: So many. I almost became a pilot and got my private pilot’s licence at 17 years old, but big changes in the airline industry meant this wasn’t possible. I flirted with the idea of becoming a surgeon, but ended up being more fascinated with building biomedical devices after a prostate cancer surgeon told me he was jealous of what I was working on during a surgery I viewed in his operating theatre. I scratched an itch after university regarding economics and fell in love with the field. If I had two lives I would love to be an economist applying that way of thinking about the world to solving hard problems in society.
My advice for someone just starting out: Be curious and open-minded. I didn’t plan my route into property, I followed interesting problems and each one led somewhere I couldn’t have predicted.
My motto/mantra: Ship early and often.
