AI GP Systems – a Market Entry Evaluation

Intelligence is central to what it means to be human, our distinguishing feature. Our ability to extend our own capability, with the fusion of technology, sets us apart.

Human developed Artificial Intelligence is becoming increasingly sophisticated at doing what humans can do, but more efficiently, quickly and at a lower cost.

I watched a BBC Horizon programme recently which looks at Babylon Health. They are actively using AI to solve an increasing range of healthcare challenges. They are activley providing GP services across London, for example, with a hybrid GP + AI service, which is proving to be popular.

How attractive would it be to enter into this market? Specifically, to build a software product which uses machine learning to gather symptoms from a patient and converting them into a diagnosis (or at least aids a GP to come to a diagnosis)? I’ve used a basic market evaluation model, as described in Personal MBA, as a way of quickly evaluating a market. I’ve done this without any research, but rather as an instinctive evaluation of the factors.

1. Urgency: how badly do people want or need this right now? The NHS is under extream pressure, particularly on the front line. Hence I’d say that a product like this, which could help to increase quality and efficiency is urgently needed. 8/10.
2. Market Size: How many people are actively purchasing things like this? I don’t think many people are actively purchasing products like this, and there are a number of barriers (not least safety and fundamental questions about whether machines will out outsmart us!). 1/10.
3. Price Potential: my gut feel. 8/10.
4. Cost of Customer Acquisition — how easy is it to acquire a new customer? Quite difficult I think, as it would currently mean a fundamental change to existing working practices. 2/10.
5. Cost of Value-Delivery: how much would it cost to create and deliver the value offered, both in money and effort? relatively speaking, a lot of investment would need to go into delivering value like this. 2/10.
6. Uniqueness of Offer: how unique is your offer versus competing offerings in the market, and how easy is it for potential competitors to copy you? Not many people competing in this market, and relatively difficult to copy. 9/10.
7. Speed to Market — how quickly can you create something to sell? In technology terms, it would be a relatively slow route to market, not least because of the safety and regulatory aspects. 1/10.
8. Up-Front Investment: how much will you have to invest before you’re ready to sell? Quite a heavy upfront investment needed. 3/10.
9. Up-Sell Potential: are there related secondary offers that you could also present to purchasing customers? I suspect there would be a reasonable opportunity for upsell, once the base product is in place. 6/10.
10. Evergreen Potential: once the initial offer has been created, how much additional work will you have to put into it in order to continue selling? I think you’d continue to need to train the model, hence there would be quite a bit of ongoing work 4/10.

My total score for this one is 44 points / 100. According to the market evaluation model, anything below 50 isn’t worth taking on. A score of 75 or above is really worth going for. Hence, I think this is one to keep an eye on until the market matures.

Please post a comment below if you disagree with my scoring or any of my reasoning. I’d love to hear what you think.