How AI software is key to UK Plc’s bounce back
I am not alone in wanting a full economic recovery to come quickly. But will British business fluff this once-in-a-lifetime opportunity?
Crisis, what crisis? So while we are having to struggle through this hopefully final phase of lockdown, a record household spending spree that will fully regenerate our economy, after the worst of the pandemic fades away, is on the cards. And surprisingly enough, the problem for British business for once won’t be people not spending enough, but failing to maximise and optimise this fantastic, perhaps once in a generation, commercial opportunity.
The facts first, in case you think I’ve gone potty here working remotely: think tank the CEBR says the world economy is recovering at its fastest rate since 1976, while British households built up an average of £7,100 in savings as lockdowns curtailed spending, taking the country’s collective cash balance to nearly £200 billion.
That organisation expects that spending boosted by that £200 billion built up during the pandemic could lead to growth of around eight percent in 2021, while the Bank of England's chief economist said in February this year that the UK economy is like a "coiled spring" and is ready to release large amounts of "pent-up financial energy”.
We need a new wave of recovery-enhancing machine learning
However, as I said, we’re not ready, though. If we roll back six months, the entire focus was on how to survive through COVID: everything was very tactical. But now, we're starting to see firms wanting to build for Recovery and looking at more strategic growth plans.
Progressive firms are looking to make their teams more efficient and pivoting to new revenue streams. But they’re also looking at correctly forecasting what is going to be happening regarding all of their business elements, whether that be customer operations or their supply chain, for instance. If business is thinking more long-term, they will want effective ways of making more accurate commercial predictions, given the business uncertainties, and for some thinking of the longer term, to automate to get even more efficient and real-time.
So moving out of COVID is going to involve a big call on software, so BI (Business Intelligence), analytics for sure, but also Artificial Intelligence (AI) and machine learning (ML). Now I’d be the first to say, as it’s a big challenge for our industry, that not enough AI and ML enterprise projects are currently linked to a business strategy.
Gartner recently advised the market that if service differentiation is a competitive advantage, we need to consider how we transform our service experience and that technology is a crucial component. The analyst firm cites useful tools to help that include applications to transform data into insights through analytics that are used by an organisation’s data science team to build models such as predicting call volumes, identifying upsell opportunities, or identifying customers at risk to cancel or churn.
So now’s the time we all need to step forward to help. We need a new wave of Recovery-enhancing machine learning that can give retailers and consumer goods players, in particular, a helping hand, as they’re the sectors most immediately likely to benefit from that spring ‘uncoiling’ in the economy. Here’s one we are starting to work with for our customers, which won’t surprise too many industry veterans: it costs far more to find a new customer than it does to keep existing customers. In telecoms, we still have the well-known churn problem, as companies have systems that try to tell them who’s likely to leave.
But these need to be so much more sophisticated than they currently are. What about knowing who are your most valuable customers? The ones you could quietly let go, as they’re not worth the trouble? If you have decisive AI with access to the entire history of you and this customer, that was using a data-powered predictive model that could tell you if this call is worth the trouble or needs even better care than our usual high standard, wouldn’t that help?
Rules-based interventions can’t give you what you want here
If machine learning is at the heart of your customer retention strategy, a brand would know that when that customer phones into the call centre or via any other channel, you do something different to the ones with a high chance of churning: you immediately move them somewhere else, which means they get immediate help, say. Or when they're interacting on your website, a little pop-up comes back and says, it's fantastic to see you again—we value you. At the moment, the retailers’ only weapon here is recommendations, but that’s not enough. Instead, imagine having real help at the call agent’s desktop that prompted the best intervention right then, either move you from a high-risk customer to a low-risk customer or to move you up the profitability scale?
Yes, some organisations are, of course, doing a form of this and reaping some benefit. But overall, this has been done to date in a mostly linear and minimal way. For brands to gain value, you want a wealth of different interventions tailored to that specific individual customer, based on their profile, characteristics and needs, which you can only get with machine learning.
The market is slowly moving towards this model—but as stated, the challenge is that many people are still using far cruder and less flexible rule-based systems to do that, which only offers relatively simple segmentation. It's not fully optimised, not deliverable at scale, and can’t happen without human intervention. Itdoesn’t provide predictability, and as it can’t make a feedback loop, the system isn’t learning from what's gone before.
To sum up, to make the most of the coming post-COVID UK and indeed cross-European retail boom, automation of decisioning based not on inflexible business rules but real intelligence will help supercharge the retail market—plus make it a genuinely data-driven, efficient and rewarding ecosystem that benefits customers as much, if not more, than providers.
John leads the EMEA technical and data science teams at H2O.ai, the leading AI technology company, and advises organisations on how to effectively apply and embed artificial intelligence into their business decision-making. Computers, numbers, statistics and problem solving have always been his passions, and he has spent his entire career...