The challenge that many of us have with “AI” (Artificial Intelligence) is that it means so many different things to so many different people. That’s not made easier by the fact that most vendors go straight into the technical details, rather than talking about what actually is the business value here, what are the different components of it, and most importantly of all: why should I care?
I think customers have ended up incredibly confused about what AI actually is, and what it can do, due to all this. For a start, 99 per cent of the time when you see the term “AI”, you should really do a mental substitution and read it as “Machine Learning” instead. Though used interchangeably, they are actually very different things.
Machine learning is the acquisition of knowledge or skill through data where AI utilises that knowledge for decision making.
And what I am beginning to find more than terminological inexactitude is that it’s becoming clearer that “AI” has to have a purpose, or it’s just not relevant. Due to Covid-19’s impact, organisations have to look at all of their business processes and consider how they can rapidly automate them, to ensure that our companies and people are productive again. Patently, this is not just an abstract conversation any more - this is existential, as it’s about whether or not you’ll actually be around for Recovery later in 2020.
IT has to step forward in this crisis to help the company get more automated and therefore more intelligent. But software isn’t the issue here. Too many companies are operating now with out of date business models. A business model is defined by our faithful friend Wikipedia “as a plan for the successful operation of a business, identifying sources of revenue, the intended customer base, products, and details of financing”.
Like it or not, everything’s being ‘Reset’
The problem is that too many (maybe all?) the business models we’ve been working with have been built over the last five or ten years, but which are now largely invalid, because Covid has changed the game completely. McKinsey has been one of the most convincing of voices when it comes to talking about the need for business leaders to wake up and smell the coffee here: “Even when Lockdown restrictions begin to ease, businesses will need to figure out how to operate in new ways, as Covid-19 could end up dwarfing the financial crisis in economic damage. In that case, it will not be enough for many companies to tweak their business model; instead, they will need to rethink it.”
In another field, HR and Learning & Development, we have been given the concept of this time as “The Great Reset” by ex-Deloitte Influencer Josh Bersin, during which, “New tools, rules, and norms will be established, and these will be with us for decades. The economy is taking a nosedive, your budget is going to get cut; this may mean a smaller team, but it definitely means redefining what you do”.
That redefinition work of your work plans, to aid the successful conduct of your business has to be very quickly updated to absorb all of the new data that's being generated by Covid and what we’re all having to do, which includes of course the constantly updating of safety advice from the Government.
The pandemic is completely changing the dynamics of customer interactions
A model can be generated in two ways; you can either have a human defining a set of business rules to create a prediction - if x happens and maybe y happens, expect this result; if p and q, then something else. Examples here could be a credit risk model that automatically works out which customers are going to default on a loan, or how you have been modelling fraud, and looking for which transactions are fraudulent, or a supply chain forecast model that used to be brilliant at predicting how many products you were going to sell on any given day at any given store or at any given location. But the Pandemic is completely changing the dynamics of customer interactions, and the supply chain, and at least some of that just isn’t ever coming back.
Given this, all Machine Learning really offers is a very scalable, accurate and powerful way to ingest your data to give a prediction of what's going to happen given certain scenarios. What great “AI” software is for, or should be for, is accelerating the process of building the new models you need NOW to allow you to generate such predictions - and which can allow the business to focus on making sure those high value models are as accurate as possible, while automating the simpler or more routine stuff, too.
But what AI is going to give you, if you clear through the fog and see it for its real contribution, is to create a new computer system based on rules that have been learnt by the computer using the latest data . Why’s that better: your new Machine Learning business models will be automatically updated to adapt to new dynamics and interactions. They’ll be scalable. And they’re going to operate at the speed you need it to now, which is as fast as not humanly, but machine-ably, possible.
Looking beyond labels
That’s why AI matters. There’s evidence that AI may even play a vital part in us getting to a cure much quicker than we’d have been able to otherwise. At my company, we’ve been closely tracking how data and algorithmic approaches have, for the first time, come together in a meaningful way to aid in informed decision-making regarding humanity’s response to this threat at massive scale. There’s also this great BBC Report, which tells us that “AI remains one of our strongest paths to achieve a perceptible solution,” but that we need to collect data at scale, which we at H2O.ai are pitching in to help as much as we can.
In any case, AI as a way to update your business model just might help you save the business. And if that’s true, do you really care what the label says?
John Spooner, head of Artificial Intelligence, EMEA, H2O.ai