The weekly grocery shop can be an hour of comparing prices, checking labels and remembering what the household needs. Richard Rome thinks AI could eventually take on much of that work, but says the industry should resist calling the result “humanless shopping”.
“But we see it being very much about humans, so rather than calling it humanless shopping, we would say it’s more human-centric with less human touch.”
Rome, associate director, consumer industry at EY, has been examining whether shoppers would allow an AI assistant to plan, compare and eventually buy groceries on their behalf. EY’s report, What Happens When Shoppers No Longer Do the Shopping?, asks what happens to retailers and suppliers when someone else starts choosing what goes into the basket.
A retailer-agnostic assistant that can do that across supermarkets is not yet available at scale. But the research suggests shoppers can already see its appeal.
Interest comes before autonomy
Sixty-one per cent of respondents have used AI to help with grocery decisions. Almost half said they would be likely to use a tool that reliably sourced their usual basket across several supermarkets, although only 17 per cent were very likely to do so.
“Not many people are ready to outsource today. That is not surprising given there is nothing to outsource it to right now; it is not really an option,” Rome says.
Asking an AI tool for a recipe or product suggestion is a smaller step than letting it place an order. Yet even limited use could change where purchasing decisions begin.
“That’s fascinating because if you think about the industry and how it’s geared up, it’s about shelf space. Digital marketing is all about search engine optimisation. This turns all of that on its head, and so it’s actually the AI agent of choice – Copilot, ChatGPT, etc, that they are getting advice from.”
For a supplier, winning a shopper’s attention may increasingly depend on whether an assistant can find, understand and recommend its products.
What would make shoppers hand over the list?
EY’s research found that saving money was the strongest incentive. About 40 per cent of respondents expect an AI shopping agent to reduce their shopping bill before they would delegate the task.
“The number one driver by far was saving money,” Rome says.
Time presents a more plausible route to adoption: 58 per cent would consider handing over an online shop if AI saved them up to 30 minutes a week, 47 per cent for an in-store shop.
“That’s probably the most obvious path to value for shoppers, and it’s quite easy to see how I could save time because you know most people are spending about an hour a week doing this for their large shop.”
The benefit may extend beyond the minutes saved. An assistant that remembers household staples, checks ingredients and works around dietary needs could also remove some of the planning that precedes a shop. A quarter of respondents saw value in help with labels and ingredients, healthier choices, dietary requirements or sustainability and ethical considerations.
That is why Rome sees an opening in the ordinary, repetitive work of grocery shopping. People need a reason to trust an assistant with their choices, but they do not necessarily need to enjoy making every choice themselves.
The households most likely to try it
Families with children under 18 stood out in the research: 70 per cent fell into the potential early-adopter group. The figures were also high for 18-to-34-year-olds, higher-income households, and people working full-time or for themselves.
“If you were to characterise it, at the top of the tree, you have time-poor families that may be carrying heavy mental loads, followed by young people that are digitally more mature or advanced, followed by people with income who are time poor.”
For retailers, that suggests the first useful applications may be those that make a familiar shop easier to complete, rather than trying to persuade customers to surrender every decision at once. Product information will need to be clear enough for an assistant to assess ingredients, claims, pack sizes and prices against a household’s stated preferences.
Who earns the shopper’s trust?
Rome believes the independent shopping assistant consumers ultimately trust is unlikely to come from a supermarket.
“For this to win consumer trust, it really needs to be centred on the shopper, and it needs to be retailer agnostic. It cannot be a Coles or Woolworths because that just cuts down the shopping field.”
He sees possible entrants in comparison services, payment providers, meal-planning platforms, marketplaces and wellness apps, as well as independent AI products. Whoever succeeds would have to convince shoppers that its recommendations serve their interests, particularly when commercial partnerships or paid placement are involved.
That creates a question for grocery businesses long before autonomous shopping becomes widespread: How will their products be assessed when an assistant, rather than a shopper browsing a shelf or app, makes the comparison?
For Rome, grocery could be only the beginning. “Grocery seems like the logical entry point. And then once it has earned trust, it will gain permission to help out in other categories.”
- Richard Rome will co-host a webinar sponsored by the Australian Food & Grocery Council on October 28. Full details here.
