Babylon and the Business World

So I identify mainly as a paediatrician. I went to Southampton Medical School, started when I was just 18, graduated at 22, and very much enjoyed being a doctor. Quick story before that: I'm second-generation Asian, we grew up above a corner shop — a fairly traditional story. I always loved the idea of getting into medicine and being able to fulfil that goal.

At medical school I quickly knew I wanted to go into paediatrics. I've got a little brother who's eleven years younger, I quite like being around kids, and I just loved the vibe on paediatric wards, where you're all aligned about getting patients home, there's high turnover, and you can have water fights. So — where better to spend your day?

After I graduated from Southampton I did my house job in Poole, and enjoyed that. Then I was thinking, well, where do I go next? I ended up coming up towards London, and without going through every step, I landed at St Mary's, part of Imperial. I found myself very luckily in a group of doctors — some more senior than me, but also the cohort I was in — who have now gone on to do some great things. There was this real vibe of, let's see what we can do to really embed ourselves in the future of paediatrics.

I remember being part of that and having the classic imposter syndrome: I don't really deserve to be here, I didn't graduate from a London university, am I really good enough for this? And what occurred to me was that where I could offer most value wasn't being the academic going off to do research on paediatric TB or HIV — I was doing the infectious diseases job at St Mary's at the time — but being quite pragmatic and getting what the service needed. I remember my clinical supervisor saying, you know what, Umang, one of the problems I've got is that I'm one of the world's best experts at paediatric HIV, but I have no idea how to get what we need when I go up against the orthopaedic surgeons who can write business cases and get stuff done.

And I remember thinking, well, maybe that's where I can add the most value — being somebody who understands how the service runs and what we need to do. Long story short, that kicked off a bit of a leadership career. I wondered, how can I go and get more skills around that? St Mary's, and especially that crew, were very supportive. That led me to a leadership fellowship with the national leadership academy, which was great. I got my postgrad certificate in medical leadership with the King's Fund and the Universities of Birmingham and Manchester.

One thing that struck me, having completed that accreditation, was that it was still very academic. What I really craved was going back and becoming something more impactful in the system. So I spent a lot of time talking to chief executives of NHS hospitals, and broadly they'd say: look, Umang, we love your enthusiasm, but money's our burning platform. We need more accountants, not more clinicians running services.

I thought that was quite a sad state of affairs. My working hypothesis was that it would be easier for me to learn how to use Excel than it would be for them to learn how to be an SHO in A&E at 4am. So I thought long and hard about where I should go to understand those skills, and I ended up at Aviva, the insurance company. Where better to learn how money flows through a healthcare economy? The NHS is incredibly complicated, but the health insurance side served a smaller population, so maybe I could start building up a basis for how we could improve things.

And brilliantly — and I have to credit my wife on this — she said, you love being a paediatrician, so before you throw it all away and go off to try something different, why don't you ask if you can keep a day a week being a doctor? And I said, no, you can't do that. Nobody does that.

And she said, look, just ask. You like doing it. What's the worst that can happen? So I spoke to Aviva before I got the job and said, would that be okay? And brilliantly, my boss said: to be honest, we just care about whether you get the job done. If you don't, we're fine — so do what you want. Which was this brilliant private-sector way of saying, make sure you get the job done. So I kept a day a week working at Frimley Park, where the clinical team were very supportive of me coming in. That allowed me to do this four-day, one-day split.

Early on in my Aviva career, I met Ali Parsa, who founded Babylon. I did my personal pitch — my elevator pitch about who I was and what I was about — and he said, you know what, when you're bored of working for an insurance company, you should come and join me. Wouldn't it be cool if we just got on with it and used technology to make things better? I'm building a small team around this yet-to-be-announced app and service called Babylon. I think it'd be a lot of fun. Let's keep in touch. And then — here I am, five and a half, six years later.

When I reflect back, it's one of my favourite teams to have been part of. Objectively for this podcast — and I don't necessarily mean he's the most successful — Bob Klaber, who's now on the management board at Imperial, was my registrar when I was a junior. Other consultants who were there, people like Mando Watson. For people not listening to this in London, these names won't mean much, but they really led services that changed paediatrics in London.

What they gave me was this ability to think that you could get involved with things that weren't just service delivery. It was all about saying, we've got all these problems — how do we as a group challenge them? Hermione Lyall was a clinical director I mentioned earlier. There was a good senior quorum. And I remember one handover where we'd got a referral from somebody else.

I won't name the people involved, but the culture was: something bad had happened, and the registrar who was on felt empowered enough to pick up his phone and go, well, I know that guy, I'll just phone him. It was basically about another doctor being a bit crappy to one of our team in a referral — which, hopefully, people listening to this will have felt, but not too often. And the culture was, I'll just phone that person and point out that that wasn't the way to do it. I thought that was brilliant, because all too often we just ignore the problem — oh, maybe they're having a bad day — and we eat it. But slowly that erodes the culture you need to build in teams. I remember it so vividly: he picked up the phone and said, I'm sure you're really busy, but I just wanted to let you know this is what happened. Taking that culture of I'm going to make a change, even if it's small — that's the group we were in.

My peer group have all gone off to be way more successful than me. One of them is now a professor of something paediatric in Australia. I feel very lucky to have been part of that group, and in many ways the beginning of my career was trying to work out how I fit within it. Culturally it was very much: let's get on and try to solve these problems. One of my fears about medicine today is that, especially when we're so busy delivering services, we sometimes feel quite disempowered. If we can create that culture of empowerment amongst ourselves, then generally we can do great things. But when we feel disempowered, it's quite hard to rise out of that and make change happen.

Yeah, that's interesting. There was definitely something about the group you're brought into, and I don't know if medicine is unique, but you do build such strong relationships, and almost everybody goes on to do great things. I really enjoy keeping in touch with people from before and knowing they can just phone up. I got a text message from somebody I hadn't spoken to in at least a decade saying, I've just had this idea and I thought you might be able to help me.

Holding onto that is really important, and there's something nice in paediatrics about it — we've always had a very MDT, team-based approach. I can't speak for the other specialties, but I think that's really important to healthcare, and as doctors we need to keep hold of it. A lot of what we do at Babylon is about saying, as a clinical voice, we're leading a clinical service. In fact, the first thing I said to Ali was that I didn't understand why an investment banker got to run a hospital group — he'd been running Circle. He laughed at me and said, well, that's why we need you to come and be part of the team. There's something about working towards the skills you've got, not trying to do it all, and forming the teams and support networks around you that let you achieve what you need to.

Absolutely. It's funny — when I go in on a Friday, sometimes the team tease me that this is just my hobby, because I bound in. This concept of a portfolio career… but really, I love the Fridays. I'm not emailing, it's just a difference, and it's helpful for paediatrics for all the reasons you can imagine: high turnover, you can do paediatric A&E and get stuff done. The advantages are huge and I'd definitely recommend it to anybody I've ever spoken to. Of course, the difficulty is that you always recommend what you've done, because that's what worked for you.

Two things have been really helpful. One is credibility. I can speak with confidence about being an actual doctor and understanding what the problems are, which is really important when you start taking leadership roles — people can follow you. I'm not saying you can't have that and then leave, but it's definitely been helpful to keep it. One challenge is that if you're junior and thinking about leaving, I wonder whether you've built up enough experience to have that credibility in the first place.

The second part is that staying in the healthcare setting makes everything else incredibly real. Think about what we're doing with Babylon — you've probably heard about us because we do AI and really cool stuff, and we've got all these data scientists building amazingly good products. And then on a Friday, one of the SHOs was telling me, and I said, can you pull up those results for me? And it took ages. I said, is the computer not working? And they said, the computer's working — it's just the keyboard isn't. So they were clicking into the handover sheet, cutting and pasting, getting around it by using the mouse. Having that juxtaposition makes you really realise: it's great to have all the AI, we definitely need it, but there's also some basic stuff we need to get off the ground. Being able to hold those two views is incredibly useful.

The downside is that anybody who does multiple roles can never really sandbox one thing into a single day. I don't only get to do 20% of my mandatory training — I have to do it all. So it's a commitment in time you have to find, and obviously that has trade-offs.

Yeah, there really is. I left to go part-time when I'd got to registrar level, and I operate at middle-grade level. When I was thinking about leaving, pretty much every consultant said, no, just make sure you complete your CCT and then you can do what you like. And at the risk of making my past narrative fit where I've got to now, the advantage of not operating at consultant level is that there's always that cover.

I remember as a medical student following an SHO around and saying, I'm really worried, I'm never going to know as much as you, I don't think this is for me, I'm never going to be as good as you. And he said, the trick is you've only ever got to be a really good house officer — you've only got to be vaguely better than what you're currently doing. That's a view I take. I'm not the world's best paediatrician, clearly, but what I'm able to do is play my part. And there's something about having the confidence to say, that's outside my competency, I need to be mature enough to accept that and ask for help or point people in a different direction. Having that awareness is important.

Good question. Do you know that thing that went around — what people think I do versus what I actually do? I had a slide like that I'd share when talking to medical students. Importantly, there are some core elements you learn during medicine that are vital for business: understanding problems, solving problems, and taking responsibility. So you're well grounded as a medic to move into other areas.

The differences, though, are that it's a very different mindset at times. There isn't a "fix this patient and then that's done" — projects last much longer. I'm pretty sure everybody listening will hate doing an audit, but that's basically a project: you start something and see it through, and there are different ways of balancing your time. You're also always tied to a very different objective. When you're a clinician you're more in control of that objective day to day, even if you don't feel you can steer the whole trust. In business it's about knowing where you fit into the overarching organisation.

The final big difference is culture. In the NHS, every team and hospital and GP practice has its own feel, but the baseline is pretty standard — you can locum in one hospital after another and generally fit in. Whereas if I compare Aviva to Babylon, they're drastically different, and there's something about being able to adapt to that. So my counsel is: if you're thinking about going into something, do your due diligence. It's not going to be anywhere near as easy as you think to go from one place to another. You might think, I can do that because I'm always changing teams — but you do have to do your homework, find out about the organisation, spend lots of time talking to people, so that when you go in you know where to fit and are clear about what you need to achieve.

Taxing question — I wonder if this will come back to bite me. I think medics do really well on emotional intelligence. We're generally good with clients and other people, and good at leading things.

I'd say we're not very good at general day-to-day management, because other than writing lots of lists with boxes half-coloured in next to them, we don't really get trained in it the way someone might have done on a more traditional business route. The other challenge medics often have is a slightly different view of the world: as doctors we're generally told we're great and valuable, which is very positive, but when you take that into business it can be hard to stay humble enough not to believe your own hype. There are quite a lot of people who come in with "but I'm a doctor, therefore I can" — that's something to counsel against. Individuals rarely display it overtly, but as a cohort it can be challenging for organisations. Sometimes people don't realise I'm a doctor because I've got a business suit on, and I'll hear the odd comment about "working with doctors" — that there's a baseline arrogance of, but we know we can do that.

The other problem is that because we always do the next thing and progress, I don't think clinicians are very good at understanding proper performance management. When you work for a big insurance company or a startup, you've got to justify your suit. At Babylon it's very much that we're on a mission, and everybody has to play their part. That's very different from the NHS, where as long as you generically tick all the boxes, you'll progress and move on. It's a different skillset.

When I started, there were about ten of us, give or take, and my job was: how do we start engaging hospitals? I'm a hospital clinician, so how do we go and do more secondary and tertiary care services using technology? That lasted about a day before we realised we didn't really have any users — primary care was the stepping stone. So I got up, walked around the desk, and sat next to our commercial lead, a great guy called Rabie who came from Google. I said, how do we start getting users, because I can't do the job I need to do until we've got people using the service? And since then I've sat, if you like, on that commercial side.

So my job is really about finding populations of people we can best use our technology for, and then crucially proving that it works for them. Right now I'm focused on the NHS — I manage the NHS team, where we look at what more we can do, building on what we've already done, to reach more diverse groups. You'll see stuff coming out in diabetes, integrations between primary and secondary care. My role is going, well, how do we flesh out what these problems are, understand where the Babylon technology and platform can best fit, make sure it gets implemented, and then hopefully showcase it back to the NHS as a use case, so we can do more of it.

Sure. It's been a fascinating experience. It almost feels like whichever way you cut it, you'll never satisfy everybody — and that's sort of the way it should be, right? Whenever you're trying to be positively disruptive… I don't think anybody, especially now we've had COVID, would deny that if we hadn't led the charge and made telemedicine more mainstream, we'd be in a worse place to have dealt with the crisis. So we're proud of what we've achieved.

But I remember, very early on, going to a meeting with some NHS people and academics. This was before we were even doing AI — purely about telemedicine. And they said to me, we haven't got any evidence to suggest we should do this. And I said, well, here's a load of evidence — here's what's happening in the US. But fundamentally they were ignoring the underlying evidence: we don't have enough doctors to deliver face-to-face appointments. Romantically, I'd love it if there were enough service to meet demand, but pragmatically there just isn't, so we have to try something.

So I said, how much evidence do you need? Give me a number. Do you need 5,000 appointments? Yeah, about that. So off it went, we got the appointments, came back, showed them the feedback — and they said, still not enough evidence. And it just went: okay, well, what do you want now, 50,000? My feeling was that we were hiding behind, I'm not quite sure where to put this, so I'll keep kicking it into the long grass and asking for more evidence. Of course, with anything brand new, you're going to have to take a bit of a leap of faith.

What we realised when we needed the AI was that the cost of a doctor behind a computer screen is the same as a doctor behind a desk — you pay for their expertise, and we're still woefully short of clinicians to serve the world's needs.

To bring our global aims to life with a story: in Rwanda, where we went live after the UK — that's our second country — there are 11 million citizens but only about 1,000 doctors. There's no way that doctor population can cover everyone, so you have to think about new ways of delivering care. To finish the Rwanda story: apparently there are 18 million mobile phones there. I was told they've got more mobile phones in the country than they have toothbrushes. I don't know how they found that out, but clearly if you're going to build services, you've got to think about how to best use what you've got to satisfy that population's health needs.

That's what led us on to doing AI. There are never going to be enough doctors, so I need to think about the best technology available today to deliver parts of that service in a way that's safe and effective. And while we're at it — safe and effective care being the tenets you can't go below — can we also make healthcare more personalised? That's the stuff you've got to see about Babylon and AI.

Then everybody said, okay, we're a bit scared about AI, so we need you to show it's safe and effective. And we said, well, we didn't quite know how to do that. We absolutely want to publish peer-reviewed journals, but it becomes a chicken-and-egg situation — I can't get the users to generate peer-reviewed evidence until we start developing and delivering the services. So what we did was take the robust processes we'd put in place to make sure what we developed was safe, and turn that into something consumable, so other people could understand our approach.

At medical school, we'd test people with cases — get actors in and say, here's a case, do your assessment. So we thought, can we do that again using AI? Can we get actors? And we did, adding as much rigour as we could before showcasing it. One challenge was that this got misinterpreted as us saying AI is better than a doctor — because that's an easy headline to write.

The point is that we don't deploy our AI outside of it being in combination with being able to see a doctor. The whole point of Babylon is that if I'm going to make healthcare more accessible, I've got to make sure the AI helps the clinical workforce get to the right patients at the right time. I think we missed the mark on explaining that well. The AI is there to augment and supplement what we do, not to say it's going to take over the entire world and we won't need doctors again — we're quite a way off that.

So there'll hopefully be more publications coming out, and we're actually really well published in AI and data-science journals. The problem is that clinicians — I can't read them, I can't make sense of the maths — but at the starting blocks we're pretty well covered on academic rigour, which is right and proper. Translating that into clinical evidence is almost like a phase-one trial before you get to full clinical trials; you've got to go through that process. You can't just say, I've had this idea, I'm going to test it in a clinical trial system.

The difficulty is that because we operate in combination with clinical services, you can't really pull out the outcomes. This is a big challenge in healthcare globally: we absolutely want to drive towards better outcomes, but it's such a complicated menu of things that impact an outcome that isolating any part in a meaningful way is always open to challenge. You can say, I did it this way, and people say, yeah, but you missed that — there are so many confounding factors.

Our view now is relatively simple: we know what we're doing is being delivered at scale, and at every part of that we're trying to capture feedback. We publish our feedback scores — you can look at our GP services in London and Birmingham and see what patients say. We've got information going back to the NHS about whether our patients end up needing hospitals less, and some good green shoots of evidence are coming through. The Ipsos MORI report that came out last year showcased that for some of our London population, so we know we're on the right track.

Our view is that eventually there'll be a comparable body of evidence that somebody independent can compare to another similar body of evidence, to prove this is the way forward. I think it has to be independent, because if we do it ourselves, I don't think it has the rigour you'd need. And all of that was before COVID — so now, again, in a romantic world you'd have all that evidence, but in a purely pragmatic world you've got to do what you can to ensure safe, effective and personalised care is put in place quickly, which is what we've seen with COVID.

For example, I was very proud that we could go and help train clinicians in Wolverhampton who'd never done telemedicine consultations, really quickly, when COVID struck — because we do thousands of appointments a day. The debate about whether it's right to do telemedicine or not has gone away. We can now prove we're doing thousands a day, those patients are getting a good service, and they're not coming to harm. So now, can we extend that to protect the frontline and develop more sustainable services for the future?

Yeah. If there's a problem highlighted, we fix it instantly, as you would with technology — it's not like you have to send out an alert and hope all the doctors read it. Other than that, we work on six- and twelve-week sprint cycles, so every six to twelve weeks everything changes. It's very hard to compare 2018 to what we're doing right now.

In terms of validation, the ideal is that you validate the entire service as a whole, and we are gunning for that, but as we said, it's really hard to do. So the next best thing is making sure every individual component part is validated as best we can. People are always surprised when they see the safety reports that go into almost every single thing we do — they're just better than anything I've seen in general NHS services. If I change a process at Frimley, there's no way it goes through the same validation we've had to do for Babylon. And the reason is that we know we're leading the charge, so we think it's our duty to say, we've really robustly dug into this.

We have some of the resource and capability to do it — we can go back and review video consultations, communicate with patients in different ways to know they didn't come to harm, and get feedback, so you can start building up that body of evidence. The way we validate any single component — let's say a paediatric flow for COVID through the symptom-checking service — the amount of data that goes into it, the double-checking… we've got teams of people saying, here's what we need to do, and all the evidence behind why we've done it that way, and we can change any of it at any time.

I was really proud to be part of it. People don't see that — they look at a chatbot and go, that's easy, you just choose the next question. But if you could see how much effort goes into making sure that next question is safe and effective, that the wording doesn't elicit anxiety or give the wrong steer… So that's an example of point validation: a group of clinicians creates the case, then we ratify it with another group of senior clinicians before we release it to market. That's point validation, as opposed to being able to evaluate the whole service.

I'd love to say yeah, I got up really early and read loads — but the answer is I don't do any of that. There are two habits I'd point to. One: I saw Helen Bevan — I don't even know Helen Bevan, but she's this amazing NHS leader, and you should listen to everything she says. She did this great talk about defining your story: almost don't worry about what the actual story is, but make it easy for people to consume, because you can change it later. People don't want to see your thinking while you're going through it.

So I created this elevator pitch — I want to be the chief executive of an NHS trust — and when I went to talk to people, they'd go, oh, now I know what bucket to put you in, here's where I can help you. I don't know if I actually wanted to be the chief executive of an NHS trust — I definitely didn't know that back then — but it helped me frame the next step. I've always been careful to have a short narrative so people can easily put me in a bucket and hopefully get me to the next part.

The second habit was always trying to say yes to meeting people — being really proactive about networking. If you've got that small story, it pays off. It took me five years of going to almost anything I could before somebody in my network said, oh, did you see Aviva got this job, that might be useful, that might be something you're interested in. People forget that — it wasn't like I searched for jobs and this one came up. Having that network is really important, and you have to be very active at building it. My wife will say I talk too much and networking isn't a challenge for me, but you really do have to get into the mindset of putting yourself out there, experiencing things that are scary, going to events when you feel like a real imposter. So those are my two habits: work your story so it's easy for people to know what you need from them, and really go for yourself.

You need a beginning, a middle and an end. Very simply, mine would be: I identify as a paediatrician — that puts me in a category, and it's lucky being a paediatrician because most people think you're super lovely and really nice. The second part is the meat of it: what I'm really looking for. So, I'm looking to do a leadership journey for my career; I think I might want to be a chief executive of an NHS organisation; I wonder what the next steps are. And the final bit is the call to action — don't forget to ask how they can help you. So my story would be: I'm a paediatrician, I want to do this, would it be okay if you connected me to someone, or do you think there are any roles in your organisation that might fit?

Don't forget to ask for what you want. If you can put that into something a couple of minutes long, most people will give you ten minutes. What you can't do is spend nine and a half minutes explaining all your challenges about whether you should or shouldn't leave medicine. You have to work hard at it — practise in front of the mirror, practise at dinner parties. To get it down to three sentences takes years, and you've got to be comfortable with that narrative. In five years of doing it, I don't think I ever found anybody who wasn't then able to say, here's the next thing for you to do — get to that event, or go meet this person. That's where you start building out the network.

Absolutely — and it's quite scary to do, because almost everybody you talk to will say, why do you want to do that? Especially when you're a doctor: you like being a doctor, being a doctor is good, this doesn't make sense, it's a lot easier for you to just be a doctor. If you ask my parents, they'd probably still just say, oh, he's a doctor — we settled into that story, that's what he does. And I still identify with that. But my counsel would be, it's really quite hard to work on your story and you do have to practise it. There's definitely something about testing out a few stories before you commit to the one you want to go off and tell. People's time is precious, but people are desperate to help you — so make it easy for them. Use your friends and family to do the framing of that story. That would be my advice.