This as-told-to essay is based on a conversation with William Tunstall-Pedoe, 56, a founder and CEO. Amazon’s acquisition of his startup and his role at Unlikely AI have been verified by Business Insider. This piece has been edited for length and clarity.
I helped create Alexa, a product that everyone has heard of and most people have used. I’m proud of what we built.
But by 2016, it was clear that leaving Amazon, which I joined after the company acquired my startup, was the right decision. Continuing to work on Alexa would have been a very different job from building and launching startups, which I love to do.
I wanted to build something that would change the world
When I was 13, I would go to a college next to my school to use their mainframe, and since then I’ve been excited by computers and pushing the boundaries of what’s possible with software.
I studied computer science at the University of Cambridge and taught there after graduating in 1991, but I felt better suited to entrepreneurship than to academia. If you create something genuinely new in software, it can be on a billion smartphones in six months and truly change the world. That’s impact.
I set out to solve what I saw as a big problem. Internet search relied on users guessing keywords to get results, rather than asking natural questions like we learn to do as children. I imagined a world where you could have that same kind of conversation with computers, which led me to found True Knowledge in 2006.
Joining Amazon was the right decision
Initially, we tried to build a search engine that would compete with Google, which didn’t work. Then, we enabled other companies to integrate our search engine into their own products — but the larger companies didn’t. For a time, we focused on SEO.
The final pivot was building a voice assistant. We created an application called Evi, which launched in the UK in 2012, a year after Apple introduced Siri. We renamed the company from True Knowledge to Evi to match our product.
As a 30-person startup, we suddenly found ourselves competing with the world’s most valuable company. We spent much of that year talking to major players in tech about being acquired. Later in 2012, Amazon bought our company.
Courtesy of William Tunstall-Pedoe
Joining Amazon was the right decision. The company invested heavily in the city of Cambridge, where Evi was based, and turned our startup into a major Amazon office. Our voice assistant became one of the company’s biggest and most exciting secrets.
Moving from running a small startup to working inside a business with hundreds of thousands of employees, with Jeff Bezos at the top, was a big change, but I loved working there. I split my time between Amazon’s offices in Seattle and Cambridge, and enjoyed going back and forth, making things happen.
When we launched Alexa, we were taken aback by the response. It was instantly successful. Today, Alexa is a household name. I’m immensely proud of the Evi team.
I wrote a memo to decide whether to leave Amazon
Amazon is known for using six-page memos instead of PowerPoint presentations to promote clarity of thought. In 2016, I wrote one to help me decide if I should leave Amazon. In the memo, I laid out these facts: I’d delivered everything I could, the acquisition had been an unambiguous success, and so too had the product. At the time, thousands of people were working on Alexa.
After about three and a half years at Amazon, in 2016, it was time to go. I wanted to re-enter the startup world.
Startups can be better-suited to exploring unconventional ideas
It’s certainly possible to launch something new within a big organization, and there are real advantages to doing so. When we launched Alexa, it immediately appeared on the front page of Amazon.com, a level of exposure that most startups could only dream of. I expect I’ll work at a big company again at some point in my career.
But if you’re trying to do something novel or contrarian, a startup is often better suited. Within a large company, all it takes is one manager deciding that resources are better spent elsewhere for a project to die. At a startup, it’s the opposite. Even if 99 venture capitalists say no, you only need one investor to say yes to keep the project alive.
Courtesy of William Tunstall-Pedoe
After Amazon, I spent time mentoring at startup incubators such as Creative Destruction Lab. Through that, I became an active angel investor, which gave me a broad perspective of the many ways startups succeed and fail.
In 2019, I launched Unlikely AI, a deeptech startup focused on building neurosymbolic AI. The goal is to combine the powerful but sometimes incorrect machine-learning models with the world of algorithms, where computers are almost always right. The mission of the business is about making AI trustworthy and reliable.
As CEO, I’m constantly swamped. Running a startup can be stressful, but working on something truly big and ambitious is incredibly exciting.
I sometimes feel nostalgic about working inside a big organization, but I love being in the startup world. For me, leaving Amazon was the right decision. I don’t regret it.
