It’s just over eight years since Angie, Andy and I set up Faculty.
From a kitchen table in 2014, to a team of over 250 amazing colleagues today – together helping companies use AI to solve their biggest problems. It’s been quite the journey.
How did Faculty start? Well, I was at Harvard when I learned about educational programmes helping people become data scientists.
The programmes are good for companies, for the world, and for PhD students – as it is hard to transition from academia into business.
Some friends and I decided to put together an equivalent programme in London. I knew Angie (one of Faculty’s co-founders) from my PhD at UCL, and Andy (Faculty’s other co-founder) from school.
Together, we set up one of the UK’s first fellowship programmes. It’s 8 weeks and helps PhDs learn the skills needed to be data scientists. Initially, we made it free so as many students as possible could attend. Nowadays, we even pay students to support them through it, and get applications from about 10% of the UK’s physics, maths and engineering PhD students.
We’ve educated around 500 data scientists over the last 8 years – many of whom now work at companies from the Guardian to Google. Actually, our Fellowship just won a Princess Royal Training Award, which we are extremely proud of.
Beyond training the next generation of data scientists, we wanted to figure out the right way to use AI so it made a difference to people’s lives.
Over the last 8 years, we have built hundreds of AI systems and solved really important problems for customers. Everything from taking down terrorist propaganda, to mitigating train delays, and helping the NHS deal with covid.
In doing so, we’ve thought deeply about how to do applied AI well, and learned lots along the way. Our key takeaway? The most important thing you can do with AI today is use it to help make better decisions.
That brings us to Faculty today. Right now, we’re one of the leaders in a new field of technology called ‘decision intelligence’.
You might look at AI companies and put their approaches into two groups. The first take an “AI first” approach. They want to replace human judgement with algorithms. The second take a “human first” approach. They want to augment human judgement with algorithms.
We’re strong believers in the second approach, and we call this ‘decision intelligence’. It’s about empowering people to make better decisions by using artificial intelligence. We think this is the right approach for two reasons. The first is principled: we think this maintains human ethics and judgement. The second is pragmatic: we are realistic about where the technology is today.
So, how do we build decision intelligence? Well, the crucial factor is to start from an important decision, and figure out how we can use technology to help. It’s about automating simple things, then surfacing insight on complicated things. It’s about helping people understand what could happen, why, and what they could do about it.
It’s the next step beyond business intelligence, which is great but retrospective. Decision intelligence is prospective, as it looks forward. I say to customers – only half joking – that business intelligence is knowing last week’s lottery numbers, but decision intelligence is knowing next week’s lottery numbers.
Organisations recognise the value of this. They know they’re supposed to make data driven decisions, but can’t do it because they are overwhelmed by data – some of which they can’t even access.
Our technology, Frontier, helps organisations make better decisions. It helps people see what’s going on, and what is likely to happen, so that they can make decisions around it.
So what’s the end game? The last few years have shown organisations need help making better decisions. How can companies navigate a supply chain crisis? How can the NHS care for more patients with less staff? How can armed forces plan operations in the face of rapidly changing threats?
The only answer is decision intelligence. Technology that helps human decision makers look forward more effectively than before, off the back of data-derived insights. Technology that performs in the circumstances that matter. Technology built through a partnership that makes the organisation stronger.
Crucially, you have to do this safely – and by that we mean fair, private, robust and explainable. Some organisations offer a false dichotomy; you can choose either safe or effective – as if those were mutually exclusive. That is wrong. I want an aeroplane to fly to America and be safe. I don’t want to choose between them – and Faculty is inventing the technology to make this possible.
