When the going gets tough, AI gets going

    

When I run introductory workshops with retail clients to help them understand the value of AI – and more importantly what wins they should reasonably expect and where they should start – one of the things we always caution is to remember that not everything is forecastable. Well if the events of the last few months have reinforced anything, it is exactly this point.

Leaving aside for the moment the thorny question of whether retailers (and many other businesses in many other sectors) could and should have moved quicker back in late January and February when the likelihood of a global spread of pandemic was increasing exponentially on a daily basis, the fact is that right now, the sort of challenges the sector is facing are exactly ones that AI can help with. 

Covid-19 has indeed proven to be, as a brand new report by B2B Marketing puts it, ‘The Great Disruptor’. For retail this has meant all sorts of challenges, from the headline statement that today consumer spending is down 25% from pre-pandemic levels, to the minutiae of everyday operational processes that just cannot happen under lockdown. 

It is not just the consumer-facing aspects of operations that have been hit. Supply chains have been disrupted. Inventory visibility and management is for many retailers (with unknown amounts of inventories inaccessible in locked-down physical stores) a daily headache. Unprecedented strains are being put on the picking warehouse, on the delivery services. 

I am sure you will already have noted that some retailers have done better than others – and I am not just talking about Amazon. Put simply, those retailers that are digitally-savvy enough to have put data, analytics, and APIs at the heart of their operations have been able to pivot fast. In retail, we like to talk about agility, flexibility, speed – now’s the time to walk the talk. AI can help accelerate innovation and ensure retailers are better prepared to survive and thrive through an equivalent mega-disruption in future. Above all, it will give them a predictive and adaptive capacity that will arm them for an uncertain future. 

But for retailers who have yet to build AI into their operations – where should they start? 

Taking the first step on your AI journey 

“A journey of a thousand miles starts with a single step” – as the old Chinese proverb goes. As a specialist in retail AI, I completely understand that many will find their first foray into AI daunting. The possibilities are huge and the potential rewards can prove nothing less than game-changing, but that doesn’t have to make starting with AI a daunting prospect. Retailers – just like any company commencing a new journey for the first time – must ask themselves where, when and how to start? 

To say AI is big business is an understatement. IDC estimates that $35.8B was spent on AI in 2019. It is widely – and rightly – acknowledged as a catalyst for transformation in retail merchandising and supply chain processes. But challenges lie in scaling AI efforts and investments to meet the needs of ever-changing operating model and KPIs. So, what are the considerations and key opportunities for making retail AI projects a success? 

Only last week, I read a new tech blog which said, “Enterprise AI isn’t something that businesses can dip their toes into … Successfully implementing AI requires that companies think big—with huge long-term goals in mind.” With respect to the author, I wouldn’t say I agree. You can indeed start AI at a more ‘humble’ level and scale up. 

When starting this process, it is critically important to understand what people in your business care about. By which I mean: what they really care about, not just what they pay lip-service to. This can be hugely helpful to get an accurate, meaningful steer on what will be seen to drive business value, and therefore what the right bias is to support decision making. 

AI is often expected to be a ‘silver bullet’ – which is where trouble starts. This is an uninformed and unfair expectation. Focus and rigour in deciding how and where to start will pay off down the line, rather than rushing in or expecting a huge immediate win. 

What I term “the graveyard of PoCs” can occur because people start at the wrong place. For those retailers looking to use AI for the first time, using it for customer-level predictions is like jumping in at the deep end. Operational wins are a better place to start – and at Antuit.ai, we’re always keen to stress the inestimable value of getting some “quick wins”: proving the value of AI in one area lets it create its own organisational ‘pull’. 

What sort of operational wins go down well? Typically, merchandising and supply chains are fertile areas for these. Areas where, you as a retailer with lots of data, inevitably are the “low hanging fruit”. This then allows Antuit.ai to wrap our models around clients’ data. 

Here are four thoughts for you to take away: 

  • Start with something simple 
  • Do not think “AI vs Humans” think “Human + AI” 
  • Evaluate what AI means to your processes 
  • Demonstrate measurable value in one or more areas before rolling out 

The next step 

At Antuit.ai, we have proven experience in making AI accessible and profitable for retailers – removing the fear and complexity many (quite reasonably) feel about taking the leap. We have developed and refined a half-day engagement workshop that helps your teams bust some myths around AI while bringing into focus exactly what it can do for you. If you would like to hear more about these workshops, or discuss anything retail AI-related, I’d love to hear from you. Drop me a line.