How to prepare your business for artificial intelligence
As the use of artificial intelligence (AI) becomes commonplace, ethical and operational concerns persist, leaving the grocery industry to navigate both the promises and the pitfalls associated with it.
Canadian Grocer recently spoke with Elliot Morris, EY Canada grocery leader and Imran Ullah, EY Canada national cloud strategy leader, about the benefits of AI, how to introduce it to skeptical employees and how to harness the technology responsibly. This interview has been edited for clarity and length.
How can artificial intelligence help grocers respond to operational challenges and shifting customer demands?
Morris: One way is around product and supply chain enhancement, and the other is around enhancing consumer experience. From a grocery perspective, the places we see it coming to life – and where there’s the most demand for it – is in forecasting, route optimization, reducing waste or creating a sustainable supply chain, and then, overall supply chain optimization. From a customer experience perspective, there is a lot going on around personalization. So, being able to target and then appeal to customers on a more individual basis, based on either their profile or on specific pieces of information has been big. And, customer service, more broadly.
READ: Canadian execs anticipate need to reskill workforce in era of AI
How are grocers currently using artificial intelligence?
Morris: We’re still in very early stages. I think everyone’s excited by, let’s call it market-facing activations. The use cases I’ve seen are in a couple areas including personalization. I’ll give a specific example: there is a product listing writer that EY has that automatically creates product descriptions for e-commerce. [AI] takes a fraction of the time creating those descriptions, which is otherwise a very manual process. But, more importantly, the tool uses product reviews to build the descriptions. It helps represent the voice of the customer in a more dynamic way and adjusts what the product description is, based on customer reviews.
How should companies prepare to introduce artificial intelligence to employees who might be wary of it?
Ullah: Some companies leverage what we call a centre of excellence model. They create a team that’s not just IT-focused, but has key [players] from each part of the business to create an education plan. The benefit of that is, as you’re educating and you’re sharing more information with the organization, the fear of leveraging these tools will diminish and hopefully the adoption rate will increase.
Morris: Disproportionately, the use cases that exist today are ones that help employees get to better answers, faster. Of course, it takes organizational support and organizational preparation and strong strategy and preparation but, at the end of the day, it does help employees get to more sophisticated answers that drive more impact in many cases. The experiences I’ve had with employee resistance doesn’t look any different than any other time you introduce a new tool to the workplace.
READ: Canadians' conflicting perspectives on AI in the food industry
What about frontline workers who are worried it will take their jobs away?
Morris: I don’t want to speak for individuals but, in many cases, AI is being broadly applied against places where people don’t want to do the work, where the work is either boring or very low skill. And in many of those cases, it’s difficult to find labour to do that type of work.
Ethical concerns surrounding AI, particularly around generative AI, have emerged. How can a company ensure the technology is used responsibly?
Ullah: An organization, as part of an AI or tech strategy, needs to develop a proper governance model to know what its customer privacy policies are, because those will naturally extend to whatever [AI] tooling the organization is using. So, understanding how those AI tools will be compliant within the constraints of that company and ensuring the right level of privacy… I think that’s where most organizations will start, by extending their governance model into leveraging these third-party tools.
How will jobs in the grocery industry transform or evolve because of wider AI adoption?
Morris: I think the first one that Imran spoke to is there being a real increased demand and need for AI-type skillsets. Whether that's retraining within the existing skillsets, and/or hiring from different talent pools altogether, I think that this adds in a whole new angle to the type of folks who can and will be attracted to working in grocery. And so, from a talent perspective, it may be akin to what was going on 10 to 15 years ago as we thought about the emergence of digital analytics within grocery, that made a big difference in terms of how the business and how the industry was run.
How do grocers determine which AI technology is right for them?
Ullah: What we're seeing and what we're advising our clients on is starting with the organization's tech strategy and seeing if they've updated it to have a proper AI approach and methodology. This aligns the company's business drivers and ambitions to what their overall tech goals needs to be. Before getting into a tooling discussion, [the organization has to develop] the right kind of use cases where it thinks AI can help tactically.
Another dimension is making sure the organization has a governance model around data to ensure the right level of control and securities are in place. With AI tools, when you're using large language models, you really do need to think through the ethical considerations that an organization will be using that data for. And making sure it's aligned to the customer privacy policies that are in place.
READ: From marketing to design, brands adopt AI tools despite risk
Companies should also look at skillset and understanding. Do they have deep technical resources that can leverage AI tools, or is that something that they'll need to look for elsewhere? Do they have partners within their vendor ecosystem that have AI skillsets and tools, to be able to collaborate with them to build some of these services? And just like cloud in the last year to 18 months, AI is going to be a competitive landscape for those roles.
What steps can a company take to ensure AI outputs align with its brand image or corporate values?
Morris: You're going to want to be able to make sure that AI is reflective of all the different dimensions of your business. And that does take doing some elements of this for yourself to make sure you're able to set the parameters and define what it is you're trying to achieve. Because I think AI will be pervasive, it will be important, it will touch all aspects of your business, and you need to make sure that's reflective of where you want to go, from a strategic perspective.
What’s one thing our readers should know about AI?
Morris: Keep it simple and focus on what you can do today. You have transaction data, image data and text data, and you want to be able to keep it simple. I worry a little bit that in the complicated nature of the debate and questions around AI, that some of the basic business principles people have become comfortable with go by the wayside. So, it's critical that people ensure they have a strong strategy. It's critical that people continue to look at specific use cases. It's critical they make sure the basics are taken care of in terms of data quality, in terms of their technological capabilities. And I think that that will serve them well when it comes to particularly difficult or more complicated, and/or new applications and approaches with AI.
How do companies ensure AI doesn’t become a vehicle for taking shortcuts?
Morris: From my experience with clients, it‘s not about taking shortcuts. It requires a lot of meaningful work to be able to use AI and optimize the use of AI. If you think about de-siloing operating models or building durable AI capability, creating robust data foundations and then creating scalable AI solutions, all that takes a reasonable amount of work. To me, it isn’t so much about short-cutting in terms of expense or in terms of time, it’s around optimizing for impact and results. To do that right actually takes effort.
READ: Canadians have mixed feelings on AI in grocery: Survey
What are the potential data privacy risks where AI is concerned?
Morris: I think people are reasonably wary of giving up their personal information, even if they like the benefits. So, as we think about adoption of AI by grocers, as useful as customer-facing AI is going to be, my instinct is that between the data privacy concerns – both real and perceived – plus making sure the user interface is seamless, grocers will lean towards more back office- or supply chain-oriented or store ops-oriented use cases for AI. Customer adoption is going to be a little bit slower on a broad base. Obviously, there's going to be specific use cases where there doesn't need to be an exchange of data to make AI work, and in those cases, like a chatbot for customer service, you could imagine a broader adoption of AI.
A version of this article first appeared in Canadian Grocer’s September/October 2023 issue.
