AI: All talk and no action?
Rob Shaw, CEO of Jaywing, a data science-led UK agency, specialising in the fields of marketing, customer management and credit and fraud risk regulation and management, shares his thoughts on the future of artificial intelligence in 2019.
For me, 2018 was always going to be the year that saw much more happening in the adoption of artificial intelligence. So, as I sit here at the beginning of 2019, it’s a good time to reflect on what happened last year and take a look at what 2019 holds.
The field of AI became far more mainstream over 2018, with more agencies, suppliers, innovators and, more importantly, users than ever before. We just need to look at take-up in the domestic market. Sonos Beam, Apple HomePod, Google Max and further Alexa skills developments have all hit the market in the last year and penetration, especially in the later part of 2018, started to move beyond the early adopters. To my mind, this is good news for industry as AI becoming the norm helps break down barriers. It also means that the way we search for and find products, services and brands is changing.
Our experience of AI take-up commercially was a bit more mixed throughout 2018, which mirrors the findings of McKinsey’s November 2018 survey of AI adoption. Aside from Brexit uncertainty making clients more cautious generally, factors such as a lack of understanding (especially at senior levels), brand safety concerns, tight budgets that mean taking a punt could be personally costly, and no head room to think, have all contributed to some clients continuing to take traditional routes. And that’s fine - we can support them in this, while taking them on a journey towards AI one step at a time.
For example, we’ve done this with numerous proof of concepts of our credit scoring AI in 2018. The results have been excellent, saving £millions on bad debt charges and improving lenders’ ability to drive more value from their portfolios, so 2019 will most certainly see greater ongoing adoption here. To be fair, the very nature of AI has previously prevented its use in heavily regulated industries where predictive models need to be explainable, and the mathematical challenge of delivering explainable AI is far from trivial. This is where AI advancements are truly excelling, having been a little later to the AI party.
But I do worry that the AI industry isn’t doing enough, either to help those more nervous clients take the plunge, or to help them understand what they’re buying. It often means asking them to be more digitally transformed than they are, which can be unhelpful; or giving them some control, which means considerably more advanced innovations that in effect “constrain” the AI, requiring serious skill. There’s no need to bemoan businesses who don’t “get it” or understand AI, because this only serves to disengage and alienate. Helping them work out an AI strategy and assess which AI is for them is a far more useful pursuit. One AI can seem to be much like another to the lay person and digging beneath the marketing is not for the faint-hearted – there’s a lot of hairy maths and methodology to understand.
But baby steps really must be taken by all who can, and an overarching AI strategy, rather than point solutions, would be even better.
Early adoption can bring huge benefits, as well as risks, and this could mean some seriously competitive commercial results for those following an AI strategy. 2019 may well see some of those who are “waiting to see” falling too far behind to catch up again, which, coupled with the issues we’re seeing in the UK around the aforementioned Brexit uncertainty and the woes of the high street, may be catastrophic for some businesses.
For retailers, not responding to changing behaviours in the use of AI through our personal devices is a bit like not getting an eCommerce strategy fast enough, leaving some businesses with big overheads and an inability to respond to the market. Only it’s happening more quickly. Of course, there were plenty of dot com failures back in the late 90s and early 2000s, mostly due to a lack of strategy; so now, as it was then, just “getting some AI” isn’t going to cut it. And it’s not just retail that will feel the pinch of competitors who are smarter, more accessible and more efficient; making well-chosen inroads to AI while developing a coherent strategy to exploit the technology would serve most businesses well.
So, was 2018 the year when AI became less talk and more action? I think it was. But there is much to be done to help companies that aren’t yet comfortable with the idea of AI to get comfortable. And there’s still plenty of room for real people in today’s business world; in fact, finding them that head room to think would be really helpful.
2019 needs to, and I think will, bring more creative innovation beyond the realms of the algorithms that have in effect been around for decades, more support from the industry to encourage adoption, and for business leaders, a focus on getting the knowledge they need to take the AI bull by the horns.
Like what you see! Signup for our weekly newsletter