AI and retail: past, present, future
The history of retail is, in one sense, the history of data technologies. From the ancient clay tablets first used to tally grain, to standardised account ledgers, to digitised computer spreadsheets; commerce has driven the development of technologies that enable the collection and use of data. And as each new technology has emerged, businesses have adopted it across the board.
There are two important points to draw from this:
- Business decisions made with data are always better than those made without. We know this because there are no surviving businesses that didn’t adopt, for example, writing. They all got outcompeted.
- While new data technologies emerge, the core capabilities of businesses, and the customer needs they serve, remain the same. Bakers in ancient Sumeria, for example, didn’t suddenly all become clay tablet manufacturers; they simply integrated the use of tablets into their baking business, and thereby started making better decisions.
AI is the most recent data technology in this lineage, and somewhat similar patterns can be expected. The biggest risk businesses face today however, in relation to AI, is thinking of it as somehow separate, or worse, still not relevant to their main operations. This attitude can be characterised as, “Oh, AI is that misty thing that does self-driving cars and robots that fall in love.” It is in fact much broader and far less mysterious.
It’s about serving customers
AI actually follows naturally from digitisation. Digital processes have massively increased the scale, speed and availability of data that businesses collect, and this in turn has created the need for automated processes to be able to use all this new data, and generate better decisions. The decisions are themselves commonly the ones that businesses were already dealing with — albeit suboptimally — and the AI improvements are neither rabbit-out-of-the-hat solutions, nor futuristic products, but specific, tangible, business-positive optimisations. These optimisations feedback into the business, and ultimately to the customer, because being better, faster and cheaper for the customer is always how data-enabled businesses outcompete the rest.
To sketch a few specifics, AI may enable a business to be:
- better by anticipating customer needs more accurately, and thereby offering the right products
- faster by improving logistics and stock management, and thereby accelerating delivery
- cheaper by reducing waste in processes/cost, and allowing savings to be passed on
In this, the impact AI is having now, is analogous to that of “online” was ten to twenty years ago. Then the shift was from physical stores to omnichannel sales, but was likewise driven by customers taking advantage of new ways to get served better, faster and cheaper. This time the shift is an internal one, as businesses move from manual, judgement-led decisions to more data-driven, automated ones, but with the same essential effects. And similarly, the shift will apply across all retail sectors — not just tech — and be at least equally disruptive, because better-faster-cheaper are areas in which all businesses can improve and out-compete each other.
Making AI work for you
Companies that adopt AI successfully always take the question “How can we serve our customer better, faster and cheaper?” as their starting point for continuous improvement of their current business. They target a selection of the KPIs they are already using (as these are material to their business), deploy AI on the relevant data to generate better decisions, and — importantly — they define, quantify and monitor success. For example: Company A may decide it wants to reduce customer waiting times. Having assessed that each minute saved is worth €1,000, it invests €2m in AI software to optimise decisions, and achieves a saving of 3,000 minutes, thus yielding €3m in value and so a net gain of €1m. Or Company B may target customer conversion rates, and using AI to identify the most promising leads, achieves a conversion improvement of x% on y euros of product to yield a value of z.
So the key question for businesses looking to adopt AI therefore becomes, “How can we serve our customer better, faster and cheaper, and which KPIs are relevant to do so?”. This is a much less misty problem than, “Should we create a brand new AI department?” or even, “What should we ask our brand new AI department to do?” But it is hard nonetheless, and this is where SparkOptimus, as Europe’s leading consultancy with a 100% digital focus, can help. We work with boards and company leaders to understand their businesses, and consider potential AI applications in two ways. Firstly, we look at the customer needs and the processes to fulfill these. We map out where decisions are being made, and which ones are amendable to AI optimisation. We can then find the right software for the job, select the right KPIs to monitor progress, and calculate costs against anticipated, trackable benefits. And secondly, we look beyond — to other companies in the same sector, as well as other sectors where AI adoption is more advanced, and ask of both: where and what is AI being used to do, what results is it yielding, what barriers is it breaking, at what kind of cost, and what are the operational implications?
Business leaders need to embrace a more data-driven way of working, and be willing to implement the changes the data indicates. At the same time, when selecting where to focus, and how to quantify improvements, businesses have to ask themselves on a deep level what customer needs they really serve, what value they bring, and indeed what kind of business they want to be. They might also ask, given the history of non-adopters of new data technologies, whether it is even a choice. Instead, the real question is not if, but how to make AI work for you, which is exactly what SparkOptimus will help discover and implement.
Interested in working together? Have a look at our Advanced Data Analytics/AI service line and get in touch!
