The idea in a nutshell : Artificial Intelligence solutions run on big data sets. Start-ups are (almost always) very small – at least in the beginning. The data assets, huge customer numbers and seas of customer information that established companies (like banks) have represent an advantage to incumbents, not start-ups, when it comes to AI deployments. And AI deployments will be critical to success in the future.

The evidence suggests that AI will disproportionally advantage larger companies

Working now for a bigger company, one of the questions which is relevant to me is: Will AI  (Artificial Intelligence) disproportionally advantage larger companies ?

I can think of 3 reasons the answer’s ‘yes’.

  • Big companies are big – like in big data: Google and Facebook have become great at data management and AI as a result of having so much customer information available to them. Their data stores mean they can deploy and garner the benefits of Artificial Intelligence better than rivals. In fact, it seems to be the main source of Google’s competitive advantage according to some. The learning mechanisms employed in Artificial Intelligence solutions depend in part on large volumes of information to associate user intent with an appropriate response / action. Even non tech players, those who worry about disruption, for example Banks and telcos, often have a lot of data from millions of customers which can train the AI systems they want to use to replace workers. Start ups don’t have that resource.
  • Big companies have legacy technology: Legacy technology is around in every big company I have ever worked for. This legacy tech is currently seen as a disadvantage for big companies when compared to start ups. Projects to improve the legacy stuff are huge, time consuming, expensive and slow. But there are examples of projects where AI solutions ‘flick’ between these systems, still using them but solving customer problems more quickly. I’ve heard it called a swivel chair interface. That’s basically a person sitting in a swivel chair taking information from one computer system and putting it in another to resolve a customer issue. These swivel chair solutions exist now. Humans do those jobs. They don’t have to be built. If AI can act as effective ‘middleware’ between different ‘green screen’ (old) and other disparate legacy technologies – if it can replace the people involved in these swivel chair interfaces – its effect could be to undermine some of the technology advantage that smaller challengers to telco and banks currently have.
  • Big companies have more money: Apart form the two factors I’ve outlined, the obvious reason to suggest that AI might work better for big companies is simply the budgets they have available to them: The ability of large companies to invest significant resources, beyond those available to start-ups in the pursuit of a new idea does seem to suggest they could do better when they try. My experience is that start-ups spend their money far more cautiously than big companies. The other side of that is that profligate bigger companies can afford to get it wrong several times with projects without being held to account.

A caveat – some startups are AI startups

There is, of course, a difference between all start ups, and AI start ups.

Some AI start-ups will get the best AI minds and software engineers and write neural networks which perform markedly better than those which proceeded them. It is possible to start an AI company and do very well. The specialisation and expertise AI companies have mean they are probably exempt from the relative disadvantage that other start-ups face.

Who would be a start-up ?

There are obviously advantages to being small (more nimble, less admin, quicker to surface new ideas) and some advantages to being big (scale, resources, data troves.) However, all businesses are going to be digital businesses and big data businesses in the future if they’re not already.

Success in the future seems to be down at least in large part to the ability of all companies to employ the data it generates and to discern the patterns available in it. The best way to do that seems to be through AI.

Obviously, just because big companies can invest in AI, doesn’t mean they will. There are a lot of larger companies who are ignoring what’s going on in the area.

Conversely, the tech companies we all know are betting big, some might say betting it all on artificial intelligence. Google now say ‘AI first’ not mobile first.

The analogy is not entirely fair. Google and Facebook are platform companies most are not.

On top of their data assets, big companies also enjoy customer inertia. They have to get it very wrong (or, conversely, the competitor challenging them has to get it very right) before people will leave.

The fact that AI solutions work better for big companies suggests that those currently in place will have an easier job staying there. With a future consisting of at least an element of AI in everything, it might be getting harder to overthrow incumbents. Maybe banking will avoid the disruption that people forecast for it.