Artificial intelligence (AI) was the hot topic at the 2018 NRF Big Show. Big-name retailers, including eBay and 1-800-FLOWERS.COM shared success stories; academics debated the ethical safeguards that need to be built into AI programs; and everyone wants to know what will be needed for AI to reach its full potential.
Yet a recent Accenture survey of 1,200 executives found that only 3% plan to significantly increase their companies’ investment in AI skills over the next three years, despite expressed optimism about the technology’s benefits:
• 71% agreed that intelligent technology will be critical to their organization’s market differentiation;
• 61% think the share of roles requiring collaboration with AI will rise in the next three years; and
• 42% believe intelligent technologies will be behind every new innovation they implement in the next three years.
The RetailWire Brain Trust debated the apparent disconnect between enthusiasm for AI and paltry investment plans. Following are excerpts of the discussion:
Neil Saunders, Managing Director, GlobalData
Yes, it is true that AI is important. Yes, corporations should be looking at it and investing in it. However, many retailers have an enormous amount of work to do in getting the basics of their propositions and operations right. That is the focus in terms of time and capital expenditure. This means more nebulous technologies like AI take a back seat. That may lose some of the early mover advantage, but until such technology is proven to deliver upswings in revenue, it is probably understandable!
Stuart Jackson, Managing Director, Regency Analysis
The terms AI or machine learning cover a huge span of technologies, some of which are happening now, such as the algorithms that drive chatbots and AI that analyzes social network interactions, looking for patterns and reacting accordingly. Others will expand rapidly once proven, such as AI in the supply and logistics chains that can “learn” to predict stock levels across the whole of a company’s estate in real time, helping to avoid items being out of stock, or ensuring that overstocked items are put on promotion. But other stuff I just can’t see happening for a very long time, like in-store robot assistants. I just don’t think customers will like it; they come to a store for human interaction. At the moment, apart from maybe the tech and marketing teams, I can’t see a reason why retailers should spend any money or lose sleep over AI.
Zel Bianco, President/CEO, Interactive Edge
[Retailers] have two choices; wait to see what others do and then decide how to proceed or jump in now and be the innovator. First-mover advantage can make a big difference and waiting while others jump in may make it more challenging to catch up down the road. I believe that if a retailer has the resources — both in people and money — they should invest now as the lessons they will learn will allow them to leapfrog their competition. Look at who the leaders are now, and you will see the same ones are at the top of the list in AI as well as other experiments that sometimes work and sometimes do not. Fail fast, as they say.
Peter Luff, President, Ipsos Retail Performance
Accenture is oversimplifying the position! It’s not a case of writing a P.O. and it’s delivered. Hard consideration of how AI will be used is required. There are, after all, a plethora of options — some may help and some may not. Think about what the task or task set is that you want to address, then create the business case that this will show a return before leaping in. Once you have decided, think about the training of the neural networks required to develop the inference engine that can then be used by the teams.
This upfront training is where the effort goes. AI needs to be trained, it does not just “do” out of the box. Retailers should in most cases get help. It seems unlikely they will have the technologists and data scientists within their ranks to tackle this alone.
The good news from my discussions and experience on this subject with industry AI experts is that retail is in no worse or better shape than most other areas, despite what Accenture may have you believe.
Cynthia Holcomb, Founder/CEO, Prefeye — Preference Science Technologies
AI cannot be “buzzed” into being relevant just because it is AI. AI is an investment in the unknown for retailers. It challenges retailers to dive deep into systems, people and process to determine exactly why an AI-enabled solution provides a better, more brilliant outcome than current systems.
How does a retailer measure the value of an AI system? Does an investment in developing and deploying a new AI system into a workforce maximize the value of each employee, the retailer’s brand, business model and products they sell?
AI-enabled systems will require retraining, rethinking and repurposing how employees interact with customers, process and systems. It’s a big deal!
Charles Dimov, Director of Marketing, OrderDynamics
Retailers that are planning to invest in AI, DEFINITELY need to consider training their staff for it. It is probably going to permeate the biggest of retailers first, and trickle down. Fail to train you staff and think through how roles will need to change to adjust — and the implementation will fail (people find ways to subvert threats to their livelihood). The important part is to make sure staff understand how this will take away the mundane from their work, letting them focus on the more in-person, creative or interesting challenges.