The Best Way To Think About AI For Your Business

Tuesday 25 May 2021

I think it’s fair to say that AI is probably the most discussed technology around at the moment. However, it’s also probably the most misunderstood.

The truth is AI is everywhere now, and understanding what it is and how it can be used can open a whole new world of opportunities for your business.

So, what really is AI? And why is it such a big deal now? 

Not being much of an AI expert myself, I thought it best to bring in some true experts in the field to help me tackle these big questions.


Matt and Tom are old colleagues of mine and founders of Fuzzy Labs, a technology company specialising in advanced AI solutions. Together we will be sharing some valuable insights on how business owners can start to think about AI now and in the coming years. 


If you want to listen to our full conversation (it’s a cracker!), please check out episode 15 of EHE Capital’s weekly podcast “Extraordinary Entrepreneurs Together”.


What is AI?


My definition of AI is a capability that a machine has to make its own decisions based on its available information. The more information and data it has on a particular subject, the more accurate the decision is.


Although we haven’t started replacing humans with AI yet, I certainly think we’re at the point where AI can make pretty sensible decisions on quite a wide variety of topics. This can significantly help businesses and give them a little bit of a competitive edge in the market.


However, just like any complex project, an entrepreneur needs to know when to bring in the right people to help them, especially with AI. As I always say, you have got to find your WHOs. 


So, to provide a more in-depth understanding of AI, I asked the Fuzzy Labs chaps for their definition. This is what they said:


AI and machine learning 


Matt: “Essentially, AI is about making a computer think like a human. However, there tends to be a crossover between the terms AI and machine learning.


Machine learning only covers a small subset of what AI does. For example, if you wanted to teach a machine to do a particular task using a lot of example data and then ask it to learn how to generalise that task on its own, that is in the realm of machine learning.


AI as a term includes many other things PLUS machine learning, so it’s important to differentiate between the two when we talk about AI. 


Another thing to keep in mind before diving into AI is that, most of the time, AI isn’t trying to create something as intelligent as a human. People are far more interested in creating something semi-intelligent that has commercial value. For example, there is no real commercial value in creating a robotic human, but a self-driving car that can copy basic human functions is a different story. 


Something like a self-driving car combines many different AI techniques into one coherent thing that provides a lot of value.”


Replicating basic human functions 


Tom: “You can break AI capabilities into three cognitive functions that humans have:


  1. Vision 
  2. Language
  3. Decision-making 


A self-driving car has loads of those different capabilities, what we call ‘models’ in machine learning. It can use vision to view a scene, identify different objects based on what it sees, and use decision-making about what it’s going to do. It’s also reading the text on signs and interpreting them using language.


You can see examples of these three cognitive functions being replicated by machines everywhere you go these days – you use them every day in your phone!”


The fact that AI is already so ingrained in our society without us noticing is proof of how well it’s working. It’s just working behind the scenes to make our lives easier!


However, where is it going to go next? And what kind of things do we need to be prepared for? 


AI being used for good


Matt: “Apart from self-driving cars, which no doubt will play a big role in the future, AI capabilities have also extended to our supermarkets. For example, Amazon’s automated shops have started popping up in a few locations (one was just opened in London). 


You can walk into these shops, take things off the shelf and walk out without spending any time at a till (or getting arrested!). It will just charge you for the things you’ve picked up. It already knows what you’ve chosen, it knows who you are and knows how to recognise suspicious behaviour. 


This is another great example of how people are combining all sorts of different techniques from different areas of AI, and using different cognitive capabilities to create something that, as a whole, is entirely novel, with tons of commercial value!”


Although some people might have their reservations about privacy and fully integrating with technology, I think the practical implementation of AI, which makes people’s lives better, is an amazing thing!


I asked Tom if he had any other examples of how AI is being used for good:


Tom: “One really good example of how AI techniques can be used to help the world is the company DeepMind. They were acquired by Google in 2014 and are probably best known for building a computer that is unbeatable at chess.


More recently, they have taken their technology and have been able to apply it to some significant real-world problems in medicine and drug discovery, particularly in the complex world of protein folding.”


AI is becoming more and more accessible


One key lesson to learn about AI is how different technologies can converge to create innovative solutions. I asked Matt what he thought about the state of technology today: 


Matt: “AI research goes quite far back, and a lot of the ideas in play today go back decades. However, what’s made a lot of these things possible is that now you have a huge number of computing resources available in the cloud.


Now we have things called ‘pre-trained models’. These are basic AI capabilities available as commodity services, usually from the major cloud providers.


For example, if you wanted AI to read the text from a document, you don’t need to build anything new – that’s something valuable and available that you can plug in straight away.


And these services are getting more and more niche too.”


This is why entrepreneurs really need to be looking into how AI could benefit their business and find the right people to speak to about what is possible.


Start small and then scale up


I asked the Fuzzy Labs chaps for some parting words of advice for any entrepreneur looking into AI:


Matt: “Start small. Start with a small problem and a small set of data. Once you think that your idea potentially has some value, then try and plug into one of these pre-trained models to get the ball rolling. These models are available to anyone as a commodity service. 


This way, you don’t need a team of data scientists to prove your concept. 


However, proving a concept is different from productionising. A lot of investment goes into making your AI product scalable and ready for market, but if you get started, you will at least know what is possible first.”


This is something we are doing more and more at EHE Capital. We have started using simple AI capabilities to automate and scale up a lot of what we are doing.


This won’t only make our lives easier, but it will also improve the service we can give to our clients.


And seeing as we are only looking for tech-led companies to invest in, we need to practise what we preach too!


For a more in-depth discussion on EHE Capital and what AI means for you, check out episode 15 of our podcast “Extraordinary Entrepreneurs Together”. Alternatively, you can email us with your questions at



Our first offering is free personalised, curated news, both here on the website, and in a daily / weekly digest direct to your inbox.