dunnhumby
Consumer Trends Tracker
What Grocery Shoppers Want Most From AI
November, 2023
How do US consumers perceive Artificial Intelligence (AI)? How do perceptions vary among different consumer groups and where should the grocery industry be focusing its AI efforts?
dunnhumby’s Consumer Trends Tracker (CTT) – a window into the nation’s grocery consumer needs, behaviors, and perceptions – is well placed to shed a light on this topic. In this current edition (Wave Five) of the CTT, we collected data from 2,000 nationally representative consumers in August 2023, and included questions about their attitudes toward AI.
Table of Contents
Robert Hayes
Customer Strategy & Insight
SECTION 1
Executive Summary
At a time of fiscal conservatism and mounting food insecurity, shoppers – especially those below age 45 – are most interested in AI applications that help save money on their grocery shop.
- There is currently a high level of distrust with Artificial Intelligence (AI)
- One third of respondents do not trust AI at all
- Only 20% ‘mostly’ or ‘completely’ trust AI
- A comparatively higher 31% of consumers under the age of 45 trust AI
- Top concerns regarding AI are the impact on jobs, security/privacy/misinformation, and the loss of human touch/connection.
- Within grocery, the most appealing uses of AI are security, budgeting, and personalized rewards and recommendations.
- Among under 45’s, personalization and budgeting are the most appealing.
- Budgeting: Significant increases in competitive pricing and private brand needs underpin the appeal of AI to help shoppers manage their finances.
- Grocery list tools, inflation trackers, and recipe apps are already using AI in this space.
- Food insecurity continues to rise. Nearly one and three 18-34 year-olds are food insecure, despite a drop in inflation.
- Personalization: the increasing willingness of consumers to engage with loyalty programs – and the expectation for retailers to reward them – signals an opportunity to use AI for personalization to gain a competitive advantage.
- Sustainability: Among under 45’s, 56% say it’s very or extremely important that a retailer cares about food waste and has sustainable products and packaging.
- Communicating the use of AI to reduce both out-of-stocks and excess inventory will resonate with younger shoppers and improve brand perception.
SECTION 2
Trust in AI
Overall, there is currently a high level of distrust with AI. Only 20% of respondents ‘mostly’ or ‘completely’ trust AI, and at the other extreme, one third of all shoppers do not trust AI at all. In this report, we will explore the reasons behind these opinions, how they differ by demographic, and what it means for the grocery industry.
Trust in AI
Despite the currently low level of trust in the US, the 20% that ‘mostly’ or ‘completely’ trust AI is higher than other markets such as the UK, where the figure is 14%. There are also sub-sets of consumers that are more trusting of AI. The most consistent differentiator in opinion towards AI is age, with a much higher 31% of consumers under the age of 45 ‘mostly’ or ‘completely’ trusting the technology.
Trust in AI, by Age
Other groups more likely to trust AI are males (29%) vs. females (14%), families with kids at home (29%), households earning more than $100k (37%) and those living in urban areas (32%). Diving into geographical differences the most trusting states are Alabama (30%), Georgia (27%), California (27%), Washington (26%), South Carolina (25%), New York (23%), and Texas (23%). A proximity to the tech industry and/or higher education correlates here, as well as a younger age skew particularly in our Alabama sample. The least trusting states are Kentucky (8%), Indiana (9%) and Louisiana (10%).
Why does it matter if a customer trusts AI? Put simply, those that are more trusting are more likely to engage with products and services that use AI. They are more likely to perceive the use of AI as a positive, which will be reflected in the retailer or CPG’s brand perception. These consumers are in essence a grocer’s target market. On the other hand, those that that are distrusting are less likely to engage at all, or even see the use of AI as detrimental to brand perception.
Put simply, those that are more trusting [of AI] are more likely to engage with products and services that use AI.
SECTION 3
The Harms and Benefits of AI – a consumer view
With the dramatic rise in awareness of generative AI models, such as Chat GPT, there are many experts warning of the dangers of unregulated AI development, and yet others labelling it all as ‘catastrophizing’. Through the CTT, we’ve been able to summarize the unprompted thoughts of 2,000 US consumers to provide an accurate representation of popular opinion on AI.
When asked about the potential harm of AI, respondents talked about a range of topics. The most prominent however, was the impact on jobs. One 35-44 year old male from Maryland said “AI automation may lead to job losses in certain industries, affecting the job market and workforce… AI raises ethical issues, including data privacy, algorithm bias, and potential misuse of AI technologies.” The negative potential of AI, as perceived by consumers, is depicted below.
AI Harm
Topics categorized from 2,000 open text responses. Bubble size approximates number of mentions.
When asked about the potential benefits of AI, respondents talked broadly about it’s potential to ‘help’ people. This included assisting with household tasks, gathering information, and supporting more vulnerable groups. A related topic was the potential to help at work, improving efficiency and output, and tackling dangerous jobs. The final benefit with high awareness was the potential for AI in healthcare settings, e.g., robotic surgery and medical research. A 45-54 year old lady from New York said “[AI can be used…] to do things far more efficiently and more effectively. We can use it in medicine to detect and treat diseases or agriculture to produce food more sustainably”.
AI Benefits
Topics categorized from 2,000 open text responses. Bubble size approximates number of mentions
Overall, respondents were clearer on the potential negatives of AI than the benefits. While there were some mentions of AI improving the online shopping experience, and better personalization of offers and rewards (categorized under ‘online experience’), there is progress to be made in explaining how and where AI is being used to offer better value and experiences to customers in grocery.
SECTION 4
AI in the Grocery Industry
While the grocery sector may not be top-of-mind when it comes to AI, the range and impact of possible applications is very significant. We tested twelve potential uses for AI within grocery, and found that security, personalization, and budgeting were regarded as the most appealing.
Appeal of uses of AI in Grocery
(% very or extremely appealing)
As with trust, we saw a very clear age divide with younger shoppers finding AI uses in grocery more appealing. Among under 45’s the most appealing use was for personalized rewards and recommendations (58%), followed by budgeting (57%). This is an age group that are feeling financial pressures the most, and so are seeking AI to be used in the most relevant way to meet their core needs. For older shoppers, 55 and over – an age group that is otherwise distrusting or disinterested in AI – security (e.g., store surveillance, theft prevention, fraud detection) is the most appealing use (40%).
Top uses of AI in Grocery, by Age
(% very or extremely appealing)
SECTION 5
AI for Budgeting
Before diving into what perhaps is the most important of topics on AI and grocery retail, a little context may be helpful.
For the first time since April 2022, when this research began, more than one third (33.7%) of respondents said they have “cut or skipped meals because there wasn’t enough money for food in last 12 months”. This figure is up 2.5 points from April this year (31.2%) through August, which itself was higher than the 2022 figures. For families with children at home, 42% have cut or skipped meals in the last 12 months.
But one might ask, isn’t inflation dropping?
Perceived inflation is 22.1%, approximately 19 points higher than actual inflation (3.0%) as of August 2023. Since April this year, actual inflation has dropped 4.1 points (from 7.1% to 3.0%), but perceived inflation has only dropped 0.5 points. The predicted lag between real rates dropping and consumer perception is evident in this latest wave of data.
Food-at-home Inflation (%)
Food insecurity is a major reality for a portion of the US population, particularly younger people. Nearly one in three 18-34 year-olds “sometimes or often do not have enough to eat”. In an upcoming report on the reduction of SNAP benefits, we conclude that food insecurity goes beyond SNAP customers.
In the most recent set of consumer trends results, 73% of respondents said competitive pricing (i.e., the need to be cheaper than other retailers) was very or extremely important. This is a significant increase of 4% on the previous wave in April this year.
Competitive Pricing
Two related behaviors that were discussed in the previous CTT report are shopping around at different stores to find the best value and checking prices online before/during a shop. Both have continued to rise in the latest wave, particularly among under 45’s, whose data is shown in the charts below:
Customer Behavior (% most or all the time)
Data shown for those aged under 45. The ‘Total sample reference’ represents all age groups
While competitive pricing measures are a good indicator of consumers trying to stick to a budget, so too are private brand measures. Generally cheaper than name brands, private brands have become increasingly popular as a means to keep grocery spending in check, and the momentum shows no sign of slowing down.
Attributes measuring customer attachment to private brands, their need for competitive prices and variety, all increased this wave by 3 to 4 percent.
Customer Needs (% very or extremely important)
The application of AI
With the backdrop of rising food insecurity, and the self-reported need for competitive prices and private brand offerings, it is no surprise that budgeting has emerged as one of the most appealing uses of AI. Some examples include a couple that challenged Chat GPT to write a shopping list for Vons in LA for under $100. The corresponding TikTok of their experience went viral with viewers amazed that the final list was under $100. Companies such as HyperWrite AI offer tools for grocery list writing, while Wegmans reported earlier this year they are trialing smart cart technology from Shopic, to track items placed in carts, along with their prices and running total.
In Canada, the Inflation Cookbook, provided by SkipTheDishes delivery service, uses AI to track grocery prices and serve up cost-conscious recipes. Based on real-time price data, the cookbook highlights ten items with significant price reductions, and then displays seven recipes that incorporate those ingredients to make meal planning accessible and convenient.
The Inflation Cookbook, presented by Skip Express Lane (CNW Group/SkipTheDishes)
SECTION 6
AI for Personalization
More than ever, customers are looking to receive relevant money-saving offers, with 56% saying it’s very or extremely important that a retailer rewards them for shopping there, up 2% from April. Meanwhile, 68% of under 45’s say it’s very or extremely important for the retailer to have lots of visible discounts/items on sale, up 4% from the last wave.
As a result of these needs, over half of shoppers are now identifying themselves in order to redeem rewards, most or all the time they shop. In addition, 47% are redeeming coupons or deals from a store’s loyalty program, and 39% are interacting with a store’s app. Among under 45’s, the number interacting with a store’s app most of all the time, is higher still at 49%, an increase of 3% since April.
Customer Behavior (% most or all the time)
The consistent rise of these behaviors show the willingness of consumers to engage in rewards and loyalty programs, and is in contrast to measures relating to low base price, which while very important, have trended completely flat.
The combination of financial struggles, engagement with personalized rewards, and adoption of technology, presents an obvious opportunity for AI, particularly among consumers under 45 years old. As mentioned earlier in the report, personalized rewards and recommendations is the most appealing use of AI within grocery for this age group.
From dunnhumby’s latest Retailer Preference Index, we know that stores in the top quartile of personalization rankings have a 41% greater share of wallet and a 14% stronger emotional connection with customers, than those in the bottom quartile. Generative AI has the potential to add a layer of proactiveness to personalization, predicting customer needs better than ever before. In addition to tailored product recommendations, substitutions and search results, hyper-personalized promotional offers, personalized shopping guides or unique user experiences become possible. Retailers and brands should implement AI to improve their personalization as a means of competitive advantage, especially among the younger demographic.
SECTION 7
AI for Forecasting, Out-of-Stocks, Waste Management
Out of all customer needs, having the items I am looking for always in stock is the most important, with 81% of people rating this very or extremely important, up two percent from the last wave of data. Fortunately, the supply chain appears to be settling down, with only 13.1% of consumers saying their intended purchases were out-of-stock in the last month, compared to 14.7% in the previous wave, and 15.9% this time last year.
If availability is a fundamental need, sitting on the base of the grocery version of Maslow’s Hierarchy of Needs, then sustainability represents “love and belonging” further up the pyramid. CTT data suggests that for roughly half of shoppers, sustainability, in its various forms, is an important factor in choosing where to shop and what to buy. However, not all consumers are created equal, and it is those under 45 for which this is a bigger issue – yes, the same age group that show significantly more trust in AI.
Among under 45’s, 56% say it’s very or extremely important that a retailer cares about food waste and has sustainable products and packaging, while 54% need retailers to offer a wide variety of natural/organic products and products from local farmers/businesses.
Customer Needs (% very or extremely important)
Those aged under 45 are 9% more likely to select sustainable products/packaging when shopping, and 12% more likely to choose organic/natural products, when compared to the general population.
An estimated 30% or more of the world’s food supply goes to waste at both the retail level and by consumers. AI can help reduce this shocking figure, by picking up demand signals, analyzing real-time data such as weather and local events, and allocating inventory more efficiently. Businesses such as Wasteless, are using AI for dynamic pricing to ensure perishable items get cheaper closer to expiration dates.
The use of technology in supply chain management is often seen as a behind-the-scenes business decision. However, this research shows that the potential for AI in forecasting (predicting the supply demand), stock, and waste management is more than just that. It is something that personally resonates with customers, particularly those under 45. Communicating the use of AI to reduce both out-of-stocks and excess inventory (which leads to food waste), will improve the brand perception among consumers of grocery stores and CPGs.
Summary & Key Takeaways
The potential of Artificial Intelligence (good or bad) is almost limitless. Within grocery however, it is always crucial to consider the needs of the customer. In the same that way that price perception impacts where people shop, it the perception of AI, how and where it’s being used, that matters. To be more specific, it is the perception of shoppers under the age of 45 that really matters. This age cohort are more trusting of AI, more likely to engage, and if the application relates in any way to personal finances, they are the most in need. In another recent dunnhumby study, we have learned that under 45’s are the most vulnerable to rising prices and food insecurity as well.
Underlying consumer trends in pricing, private brand, promotions, loyalty and sustainability align to the most appealing uses of AI, according to respondents. These include AI for budgeting, personalized rewards and recommendations, and forecasting, stock and waste management. It's useful to contrast these to less appealing uses such as driverless technology, chatbots and health/diet goals, which take many of the headlines.
Implementing and communicating the use of AI to aid budgeting, personalization, and sustainability, will resonate with customers. Against a backdrop of financial difficulty and food insecurity, particularly for younger consumers, well-placed investment in AI can prove a genuine competitive advantage.
