AI adoption is growing—now’s the time to leverage its potential
The adoption of artificial intelligence (AI) is accelerating and opening a world of opportunities. But, there are still questions and uncertainties surrounding AI’s accuracy (its only as good as the data it learns on), ethical considerations and more. We recently chatted with Andreas Duess, CEO and co-founder of 6 Seeds Consulting, a data-enabled AI-driven marketing firm specializing in CPG and food industry innovation, about harnessing AI to build customer loyalty, the role it can play in pricing and promotions and perpetuating bias. (This interview has been edited for clarity and length.)
How do you ensure the data you’re working with is complete, current and accurate?
Well, you wouldn’t use ChatGPT, you would use data that you have collected and that you have audited. One of the things AI is good at is taking huge chunks of data and figuring out the connections within that data. If I’m a grocer, I would instantly look at the data I have available, I would structure this data as best I could and give AI access to all of them. Then I would start talking to my data.
How can AI help build customer loyalty?
If I feel you know me and do things that benefit me, I will be yours forever. I have three teenage boys that behave like a swarm of locusts. We can’t buy food fast enough because I fill the fridge in the morning and they come home from school and it’s bare, it’s barren. If you, as a retailer, understand me and come to me with, I don’t know, a newsletter or an article that says how to feed three teenage boys without bankrupting your family, I would find that very interesting indeed.
READ: How retailers like Walmart are winning with data
Will AI transform pricing and promotions?
Let’s look at people on GLP-1 [drugs like Ozempic, etc.]. In the U.S., 38% of the population will be on a GLP-1 drug in two years’ time and I suspect in Canada the numbers aren’t far behind. And if, for example, a retailer is helping their customers on this weight-loss journey, price becomes less important because they are working towards an outcome that’s important to those customers. If I’m running President’s Choice and I need to make goods high in fibre, nutrient dense, taste amazing, but they need to be affordable, I will have my ingredient database and I will ask [AI], “what can we do with healthy ingredients that don’t cost the earth?” And AI is good at helping us answer these things because AI is amazing at pattern recognition.
How should grocers be thinking about AI when it comes to merchandising and marketing?
AI goes wrong when people use it to write a marketing plan or a TikTok script. AI works well when we say, here's the human goal I wish to achieve, here is what I wish to communicate, here's what's important. And there are two words that we frequently use. One is sense-making and the other is meaning making. Meaning making is what humans are good at. Meaning making is when I start a food business and I want to honour my grandmother's heritage and her recipes and the feeling of love and connection we had when we sat around a table. Sense-making is asking AI for 10 product ideas. The answer will be mediocre and you will never have that connection to the emotional and human part that is so important. The same thing is true for a retailer. I think it's important for a retailer to understand the human element of their customers and then support that humanity through products that elevate that humanity. AI, if used correctly, can help us to elevate humanity and what makes us human.
How do you avoid bias?
The important thing to understand is AI will get you to 80% [of what you’re trying to achieve]. For example, let’s say part of my marketing mix is culinary content development. I have a recipe journal and I have a test kitchen with chefs who come up with 20 recipes a month. With AI, you can put in your relevant consumer data and ask for 500 recipes. If you publish these 500 recipes without human interaction, you’re setting yourself up for ridicule because some of them will taste like absolute crap or the measurements will be wrong. Instead, give those 500 recipes—developed by AI in half an hour—to a chef and ask them to pick the 51 recipes that speak to your customers. Use the customers as a starting point for your own humanity and your own genius. The outcome will elevate the skills of your test kitchen chefs because ... they can apply their knowledge, their passion and their own humanity. That’s where the magic of the interaction is. And that’s where you ask, is this bias? Does this assume things [that] need to be double checked?
How can smaller grocery businesses use AI to compete with larger chains?
If you’re an independent store, you can use the public models. There is always this fear the models will steal your data. No, they don’t. Your data is a fraction of a drop in the ocean that nobody is interested in. When you check how much inventory is being sold, put that data in a spreadsheet, upload the spreadsheet, wait three months, do the same thing again, wait three months, do the same thing again. What trends do you see? I talked to a retailer who uses an AI model he made to run his sales data against the weather report. He anticipates the right time to put specific products on sale ... aligning with weather patterns that suggest those products will be in demand. Really clever stuff you can do in a humble way, but will still help you.
READ: NIQ’s Hanif Mohamed on tech adoption
Are there any other emerging technologies or tools you're excited about?
I am excited about the ability of technology to help us build the stuff that speaks to us as humans. I'm convinced that people who understand what makes humans human will be the people that are best suited to succeed in the age of AI. If I was a retailer, on my team, I would foster creativity, I would foster emotional awareness, I would foster everything that makes humans, humans. And then, I would give them the AI tools that supercharge that ability. That's what I would do.
Looking ahead five to 10 years, how do you see AI evolving within the grocery sector?
To be perfectly honest, the technology is moving so fast, it’s difficult to look into the future.
One bit of advice … work with your data, do it quick and allow people to talk to your data and have people whose job it is to do crazy s**t. AI is good at the “what ifs.” Hey, what would happen if we take our coffee grounds and we start selling them as garden fertilizer? What statements could we make? What claims could we make? And looking at our consumers, who’d buy this?
This article was first published in Canadian Grocer’s February 2025 issue.
