It’s hard to avoid the discussion of AI, so what are some of the use cases for AI in Retail Media. The AI hype machine has been in overdrive, so it’s hard to work what is useful, what should be useful and what could be useful.
In fact, the biggest hype overdrive challenge is the fact that most of what is talked about in “AI” is not AI at all, but machine learning. Secondly, most writers about AI use the two letters as a shorthand for “what is ChatGPT useable for”?
None of this is helpful. What are some actual use cases for Retail Media and AI?
When dealing with large, complex datasets, AI becomes essential for processing data efficiently while maintaining high accuracy and speed, such as with audience segmentation, demand forecasting and optimising returns.
These can be categorised under a few headings:
Creative: Advertisers can quickly create a lot of creative for at a much lower cost and complexity than before with GenAI
Microsoft launched Retail Media Creative Studio in the Microsoft retail media platform (powered by PromoteIQ) in January 2024, but with the sunsetting of PromoteIQ, it remains to be seen if this will work with partner of choice, Criteo.
Amazon Ads image generation is a free generative AI solution that enables brands of all sizes to easily make lifestyle and brand-themed images based on product details and refine them through short prompts or by applying seasonal and lifestyle themes.
With all of the integrations offered by OpenAI and the power of DALL-E, it cannot belong where image generation is part of all the major retail media AdTech vendor capabilities.
Contextual Content: Generative AI can be used to dynamically change campaign content based on contextual data like location. For example, content could be changed depending on time of day, location, and weather.
A much more powerful use of context is the combining inventory data into ads: advertising an out-of-stock products wastes budget as well as being out of the shopper basket and missing a sale.
Walmart Connect and Walmart Luminate have combined advertising and product inventory to refine the content of ads, but few have managed to directly link inventory dynamically. Instead, the link is typically rule based, not AI based. However, using machine learning, you can see how this could be much more dynamic and real time.
Campaign Optimisation: Retailers can follow the lead of Meta and Google with Advantage+ and Performance Max by using AI to continuously analyse campaign performance and automatically adjust bids and targeting to maximise ROI.
Personalisation: AI-powered personalisation in retail media tailors’ content to individual preferences. Personalisation dynamically adjusts creative, copy and imagery based on shoppers like gluten-free shoppers or health-conscious shoppers, AI creates more impactful, tailored ad experiences. Brands using AI and machine learning for ad personalization have seen a 1.3x increase in ROAS, according to Kroger Precision Marketing.
Insights and Analytics: AI is excellent for summarising large quantities of data to build a broader picture – and is perfect to help drill down what the best data points are to focus on. These can be addressed across the three main types of analytics:
– Descriptive Analytics: Descriptive analytics are based on historical data, aggregating data and visualising processes to help identify patterns in datasets accurately. For example, AI can sift through vast amounts of retail media performance data, summarising trends in real time and generating automated reports.
– Predictive Analytics: Machine learning algorithms analyse past behaviours and data to forecast future outcomes. In Retail Media, AI models learn from customer interactions, purchase history, and engagement metrics to predict customer behaviour, such as which products they are likely to buy or which ads they will respond to.
– Prescriptive Analytics: AI uses optimisation algorithms and machine learning to recommend the best course of action. In retail media, AI would suggest actions, such as which ads to run, when to run them, and how to allocate budgets most effectively. Prescriptive AI models can even automate decisions, dynamically adjusting campaigns in real time based on customer behaviour and external factors
Propensity Models: Propensity models predict the likelihood of a customer taking a specific action, such as making a purchase or engaging with an ad. These models identify which shoppers are most likely to respond to certain products or promotions. Retailers can then use these insights to create highly targeted, personalised ad campaigns. Propensity models allow for a “segment of one” approach, tailoring ads to the individual level for maximum relevance and impact.
Of course, all of these modelling and approaches are nothing without the insights of humans who can draw conclusions and make recommendations that use judgement rather than just dry facts and data. However, you could argue that even this approach could be supplanted by AI.
How are the retailers applying AI and machine learning?
The recent announcements from Walmart give us some real live examples of what retailers are thinking of.
- Walmart has developed a proprietary GenAI platform called “Wallaby” a retail-specific large language model (LLM) tailored to enhance customer interactions. Walmart have trained the LLM on decades of Walmart data to delivers highly personalised, contextual responses in line with the company’s values.
- Walmart has also upgraded its AI-powered Customer Support Assistant, enabling it to instantly recognise customers and take actions like managing orders and returns. The company is rapidly expanding its GenAI tools for customers, associates, and partners, including advanced assistants for Sam’s Club and Walmart International.
- Walmart has developed a “Content Decision Platform” that serves as a tool to create shopping experiences tailored to the individual customer. The platform leverages AI-based technology to understand the customer and a GenAI-powered tool that can predict the type of content they’d like to see on the site. It’s already being used within select areas of Walmart.com.
- With this technology, Walmart will create a unique homepage for each shopper making the online shopping experience as personalised as stepping into a store designed exclusively for each customer. The U.S. version will launch by the end of 2025 (!) and the company plans to also use the platform’s underlying technology in Walmart International’s Canada and Mexico markets for personalised item recommendations
- Walmart is using GenAI to improve the quality of its product catalogue by using large language models (LLMs) to create or enhance over 850 million pieces of product data. The company said that without the LLMs it would have needed around 100 times as many people to complete the work in the same amount of time. It is also using GenAI to help improve its multilingual search functionality.
You are in: Home » Retail Media » ANALYSIS The AI hype train is pulling in at the retail media platform – what is it delivering?
ANALYSIS The AI hype train is pulling in at the retail media platform – what is it delivering?
Colin Lewis
It’s hard to avoid the discussion of AI, so what are some of the use cases for AI in Retail Media. The AI hype machine has been in overdrive, so it’s hard to work what is useful, what should be useful and what could be useful.
In fact, the biggest hype overdrive challenge is the fact that most of what is talked about in “AI” is not AI at all, but machine learning. Secondly, most writers about AI use the two letters as a shorthand for “what is ChatGPT useable for”?
None of this is helpful. What are some actual use cases for Retail Media and AI?
When dealing with large, complex datasets, AI becomes essential for processing data efficiently while maintaining high accuracy and speed, such as with audience segmentation, demand forecasting and optimising returns.
These can be categorised under a few headings:
Creative: Advertisers can quickly create a lot of creative for at a much lower cost and complexity than before with GenAI
Microsoft launched Retail Media Creative Studio in the Microsoft retail media platform (powered by PromoteIQ) in January 2024, but with the sunsetting of PromoteIQ, it remains to be seen if this will work with partner of choice, Criteo.
Amazon Ads image generation is a free generative AI solution that enables brands of all sizes to easily make lifestyle and brand-themed images based on product details and refine them through short prompts or by applying seasonal and lifestyle themes.
With all of the integrations offered by OpenAI and the power of DALL-E, it cannot belong where image generation is part of all the major retail media AdTech vendor capabilities.
Contextual Content: Generative AI can be used to dynamically change campaign content based on contextual data like location. For example, content could be changed depending on time of day, location, and weather.
A much more powerful use of context is the combining inventory data into ads: advertising an out-of-stock products wastes budget as well as being out of the shopper basket and missing a sale.
Walmart Connect and Walmart Luminate have combined advertising and product inventory to refine the content of ads, but few have managed to directly link inventory dynamically. Instead, the link is typically rule based, not AI based. However, using machine learning, you can see how this could be much more dynamic and real time.
Campaign Optimisation: Retailers can follow the lead of Meta and Google with Advantage+ and Performance Max by using AI to continuously analyse campaign performance and automatically adjust bids and targeting to maximise ROI.
Personalisation: AI-powered personalisation in retail media tailors’ content to individual preferences. Personalisation dynamically adjusts creative, copy and imagery based on shoppers like gluten-free shoppers or health-conscious shoppers, AI creates more impactful, tailored ad experiences. Brands using AI and machine learning for ad personalization have seen a 1.3x increase in ROAS, according to Kroger Precision Marketing.
Insights and Analytics: AI is excellent for summarising large quantities of data to build a broader picture – and is perfect to help drill down what the best data points are to focus on. These can be addressed across the three main types of analytics:
– Descriptive Analytics: Descriptive analytics are based on historical data, aggregating data and visualising processes to help identify patterns in datasets accurately. For example, AI can sift through vast amounts of retail media performance data, summarising trends in real time and generating automated reports.
– Predictive Analytics: Machine learning algorithms analyse past behaviours and data to forecast future outcomes. In Retail Media, AI models learn from customer interactions, purchase history, and engagement metrics to predict customer behaviour, such as which products they are likely to buy or which ads they will respond to.
– Prescriptive Analytics: AI uses optimisation algorithms and machine learning to recommend the best course of action. In retail media, AI would suggest actions, such as which ads to run, when to run them, and how to allocate budgets most effectively. Prescriptive AI models can even automate decisions, dynamically adjusting campaigns in real time based on customer behaviour and external factors
Propensity Models: Propensity models predict the likelihood of a customer taking a specific action, such as making a purchase or engaging with an ad. These models identify which shoppers are most likely to respond to certain products or promotions. Retailers can then use these insights to create highly targeted, personalised ad campaigns. Propensity models allow for a “segment of one” approach, tailoring ads to the individual level for maximum relevance and impact.
Of course, all of these modelling and approaches are nothing without the insights of humans who can draw conclusions and make recommendations that use judgement rather than just dry facts and data. However, you could argue that even this approach could be supplanted by AI.
How are the retailers applying AI and machine learning?
The recent announcements from Walmart give us some real live examples of what retailers are thinking of.
Read More
You may also like
Register for Newsletter
Receive 3 newsletters per week
Gain access to all Top500 research
Personalise your experience on IR.net