More UK retailers are starting to experiment with AI-driven analytics. A data platform that can evolve as AI use matures will be key to achieving long-term value.

Retail adoption of AI is growing- fast. A survey conducted by PWC for Retail Week in June 2019 found that 50% of UK retail firms with more than £50m turnover are using AI in their business today. More than one-fifth (21%) plan to invest between £1m and £20m on it between now and 2023.

Data and analytics are key areas for AI investment

Data and analytics are emerging as a key areas for investment, cited by 56% of respondents to PWC’s survey. Forward-thinking retailers are setting up innovation teams to explore the potential of AI-driven analytics, and assess how it could be operationalised in their business.

M&S is one example. “If you think about how retailers are operating in the modern world, the datasets being generated simply can’t be assessed or investigated by a single person,” Paul Dasan-Cutting, Innovation Manager at M&S, told Microsoft in 2019.

“We have to come up with a set of models to interpret that data more efficiently. [It]can help us offer even better products to our customers and also rethink how we plan our operations.”

A ‘technology-first’ mindset is the wrong approach

While retailers like M&S are laying solid groundwork for introducing and scaling predictive analytics, adoption across the retail industry is mostly still at the experimental stage.

In its 2019 report Accelerating Competitive Advantage with AI, Microsoft found that skills and knowledge around how to use AI are still lacking across the industry, with less than a fifth (19%) of executives saying people in their organisation can describe how AI can help the organisation to achieve its goals.

That’s leading to a ‘technology-first’ rather than ‘problem-first’ mindset, according to Alex Sbardella, Senior Vice President of Global Innovation at retail innovation consultancy GDR Creative Intelligence.

“AI projects suffer when retail leaders start with the technology and not the problem they are trying to solve,” he told Microsoft. “Any innovation project must begin with the customer or business need. If AI is deemed the right solution, then do your due diligence, get your data in order, your systems and your people ready.”

Many AI innovation projects have no roadmap for scaling

I would add a further piece of advice to Alex Sbardella’s comment: don’t over-focus on today’s project at the expense of future needs.

One risk factor we often see with AI and analytics innovation projects is that there’s no roadmap in place for scaling or evolving them as new applications, data sources and technologies emerge.

Microsoft sounded a warning to this effect in its 2019 report, saying that “[a] challenge organisations face as they seek to scale AI adoption is ensuring they’re able to manage the technology as its functions and use cases evolve.”

AI use cases are evolving fast

That’s especially true for retail, an industry that’s generating data at a phenomenal rate. Retailers are adding new data sources all the time: from in-store video feeds and smart shelves to sensor data from the supply chain, customer social media activity and external weather forecasts.

An AI pilot may focus on just one area – for example predicting the impact of the coming week’s weather on store sales of a particular category. But if the pilot is successful, it may be quickly adopted across the business.

“AI can be used across varied areas [of retail],” notes PWC, “including store layout, roll-out planning, purchasing, logistics, resourcing and recruitment.” The FT sees dynamic pricing as another imminent AI-driven trend, reporting that “by 2050 the idea that everybody pays the same amount for an item will appear bizarre.”

The data platform must evolve as AI use matures

As data science and AI become embedded across every aspect of retail operations, the ability to manage and evolve the analytics platform will be crucial to maintaining competitive advantage.

If the platform can’t easily accommodate new sources of data, retailers risk being unable to generate new business-enhancing insights. And if it can’t easily make use of new technologies as they emerge, it risks quickly falling into obsolescence.

In addition to hiring data scientists to develop predictive models and make sense of the insights, retailers will also need technology specialists to build, manage and evolve the supporting data platform.

Those specialists will need to be familiar with the state of the art in large-scale data and analytics platforms. They will also need to keep abreast of new technological developments as they emerge, so the retailer can use them to gain competitive advantage.

For most UK retailers today, those capabilities are not in place inhouse. Retailers face a choice: hire one or more data platform technologists – perhaps at the expense of hiring one or more data scientists – or work with an expert partner with the right skills, knowledge, and early access to emerging technologies.

Finding the right partners for the AI journey is one of the key takeaways of PWC’s study of AI maturity in retail. “Choose partners that have the skills your business does not,” it says, “and encourage them to invest alongside you and to help upskill your team.”

Adatis: an expert partner to manage and evolve your retail data platform

As a Microsoft Gold Partner specialising in data and analytics, Telefónica Tech works with retail businesses across the UK to build, manage and evolve data platforms that stay current as new business needs – and enabling technologies – emerge.

We give clients the flexibility to choose their own path: learn from us and then take over management of the data platform, or engage us as a long-term partner to manage and evolve the platform while they focus on building up their inhouse data science capability.

To learn more about how we can work with you to ensure your data platform is always fit for purpose and delivering competitive advantage, please get in touch.