Retailers excited about gen AI; retailers don't have the data foundations in place for gen AI - Salesforce study exposes some harsh realities

Stuart Lauchlan Profile picture for user slauchlan April 3, 2024
You can't build a revolution on dodgy foundations - retailers have a lot of work to do on data strategies if they are to realize the benefits of generative AI.


As evidenced on diginomica passim, major retailers have begun to talk openly of their ambitions around generative AI, some more pragmatically than others perhaps. But whatever the scale and scope of those ambitions may be, the common thread is that retail, like every other business sector, has a lot to think about with respect to gen AI's potential benefits and downsides. 

But, again in common with other business sectors, there are a lot of questions to answer about the how, why and wheres of adoption, as well as the thorny reality of just how prepared organizations actually are to reap the proclaimed benefits of generative AI today. 

A salutary new report from the Retail AI Council - whose members include some very big names - and Salesforce, based on a study of over 1,300 retailers around the globe, provides a timely reminder on that last point. It finds that nearly half of all respondents admit to struggling to make their data accessible, with just 42% claiming to be able to connect up their various data silos. 

As we all know by now (or ought to know!), data is the bedrock of successful generative AI adoption. Without having a handle on that, it’s hallucination time in the shopping aisles. And while the likes of OpenAI CEO Sam Altman may try to argue that hallucinations are a feature, not a bug, retailers are thankfully not buying that line. According to the study, 63% of all respondents cite bias as the top risk in using generative AI, followed by hallucinations (38%) and toxicity (35%).


That said, the interest in gen AI is definitely there. Some 81% of respondents say they have a dedicated AI budget, with an average of 50% of that assigned to generative AI. So, let’s start with some positivity. There is a lot of potential here for retailers. According to a 2023 study from analyst firm IHL Group, generative AI could have an impact of $9.2 trillion on the retail sector by as early as 2029, the three main benefits being increased sales, improved gross margins, and lowered Selling and Administrative (S&A) costs.

Nearly a third (30%) of global respondents in the Salesforce study say that they are currently executing on gen AI, while a fifth claim that they are realizing benefits and gauging further investments. Just below that on 19% are those retailers that are at the 'strategizing' stage and working on business cases. Thirteen percent say they’re 'exploring' or 'considering' generative AI, but don’t yet have a formal plan in place or any budget. Only five percent say they don’t regard AI as a priority. 

There are three main focuses of gen AI strategy today - Customer Experience (50% of respondents), Employee Experience (35%) and operational efficiencies (15%). The most important business functions to which gen AI tech can be applied are customer service (34%), marketing (30%), store operations (30%), supply chain (26%) and digital commerce (26%). 

Personalization and customer service use cases are cited as target areas for generative AI adoption by retailers. The study finds that just over a third of retail employees are using gen AI today, with that expected to rise to 45% by the end of next year. Of that percentage today, some 93% say they are already using generative AI around personalization, in the form of email copy or product recommendations to customers. That number is expected to rise to 96% within 18 months. 

This personalization can take a number of forms. Some 32% cite a use case of generating personlized responses for agents to quickly email or message customers, while 30% are using the tech to create a conversational digital shopping assistant to help shoppers find the right product or service. Other use cases include creating personalized promotion for loyal customers (27%), producing creative assets for ads, emails, social media etc (26%) and creating personalized promotional offers (25%). 


That’s the positive. The negative bring us back to that damn data issue. Drilling down from the headline conclusions, the underlying detail doesn’t get any more comforting. The study finds that only 17% of respondents reckon to have a complete, single view of their customers. Nearly half (49%) are still in the preliminary stages of building or even considering the creation of a complete customer data profile.

As for what all that means for retailers building gen AI models, while over two-thirds (67%) of retailers boast that they are fully able to capture customer data, only 39% are fully able to clean that data, with over half (54%) only able to say they are ‘somewhat able’, whatever that means in practical realities. 

Meanwhile just 42% say they are fully able to harmonize data, with 47% ‘somewhat able’ and an alarming 11% who admit they’re not at all able. That’s a lot of Altman’s ‘features’ in the making right there. 

So what’s being done to put this right? Well, it’s a mixed bag in terms of responses. Less than a third of respondents (30%) can say that they are actually executing on creating a single customer view, while only 17% say they have this in place and are realizing benefits. Some 22% are at the 'strategizing' stage, 14% say they are ‘considering’ the idea of a single customer view - what the hell is there to consider? - while a further 13% are ‘exploring’, which seems to mean keeping an eye on what others are doing in this regard. 

One final finding from the study that may alarm many. Over half (51%) of respondents reckon that their generative AI output is fully-automated and used without any human intervention, while 41% cite ‘light human intervention’ before being used. Only eight percent reckon it has full human intervention. 

Caveat emptor!

My take

I’ll leave it to Rob Garf, Salesforce’s GM of Retail and Consumer Goods, to sum up: 

The AI revolution is about data, trust, and customer experience. Looking at AI in isolation, without understanding these elements as a package, will hurt a retailer’s ability to build loyalty and improve customer relationships. 

Go sort out your data management, retailers!

A grey colored placeholder image