Threefold revolution: The influence of generative AI on retail and consumer goods
While generative AI—initially in the form of ChatGPT—may boast the steepest adoption curve in the history of technology, the scramble to use it to accelerate business value is far from over.
In just over a year, it has gone through what you might call the ‘shiny toy’ stage where teams play with it to try and work out what it can do for them. From this, lessons have been learned and applied. Some of the lessons Microsoft teams have learned have been highlighted in previous blog posts.
Microsoft’s customer teams have undertaken many customer workshops, each focused on identifying the areas that have the greatest opportunity for benefit.
McKinsey suggests that for retail and consumer goods businesses, the value potential is somewhere in the region of 1 to 2% of the total industry revenue. As for the ‘low hanging fruit’, about “75% of the value that generative AI use cases could deliver falls across four areas: Customer operations, marketing and sales, software engineering, and research and development (R&D).”1 But in practical terms what does this look like if you are a retail or consumer goods company?
From the work Microsoft has undertaken there are three broad groups of use cases that offer the greatest value:
- Content and product marketing.
- Internal knowledge management.
- Customer conversational experience.
You may wish to explore each of these areas with a view to understanding what others are doing and considering examination of something similar in your organization.
Content and product marketing
At the heart of generative AI is the ability to create new content, so it stands to reason that this would be an area of high potential.
Content marketing traditionally involves a series of iterative loops involving multiple parties—perhaps a brand or product owner and a creative group—be this a copywriter or a creative agency that produces images.
This approach has several challenges. Firstly, due to the iterative nature between different parties it can take several days or weeks of iterations due to the lags between each create—review—revise cycle. For visual content—with the use of external agencies, complex backgrounds, complex picture, and editing—this can become expensive.
These two reasons mean that scaling the creation of content becomes very difficult. If you have a very wide range of products or a wide range of customer groups for which you would like to customize the message, it is simply impossible to achieve this with a traditional approach.
Generative AI changes this.
One of the first case studies regarding the use of generative AI that Microsoft highlighted was the creation of product related marketing content at a vast scale. Carmax—a used car retailer—wanted to provide a consistent set of product information for all the different makes and model of cars that they sell. Generative AI was used to generate text for car comparisons allowing viewing of specifications, features, highlights, and summary reviews. Carmax estimated that to build this would have required eleven years of effort—which dramatically illustrates how generative AI can address the scaling challenge when a retailer has a wide range of products. Learn more about the Carmax case study alongside example content and a short video.
Marketers aspire to segment their customers into smaller and smaller groups to make messaging as personal as possible. Customer data platforms such as Microsoft Dynamics 365 Customer Insights allow creation of segments based on customer attributes from multiple sources. Websites and social media platform allow specific messages to be targeted at these groups but the challenge of having the capacity and time to create the relevant content remains a constraint.
This is where generative AI can be used to fill the gap. A number of organizations are utilizing an innovative approach of aligning keywords to their products and then using generative AI to suggest a series of advertisements, or social media headlines associated with specific consumer profiles. Following review by a copywriter, to ensure brand alignment and an appropriate tone, these headlines are then approved for use. This approach can enhance overall creativity as well as enabling more granular targeting.
Internal knowledge management
“If HP knew what HP knows, we’d be three times more productive.” This is a quote attributed to Lewis Platt who was Chief Executive Officer of HP between 1993 and 1999 and is well known amongst knowledge management professionals.2
It is no secret that organizations create and retain a lot of knowledge. The larger the organization the more knowledge. But more knowledge can often add to the problem—understanding what is available can be very difficult. As Lewis Platt suggested, organizations do not know what knowledge they have. Knowledge becomes siloed across the different systems that permeate the organization and pulling it together for specific purposes becomes very difficult.
Traditional search might be able to help you find something specific within your organization by referring to a particular document. It will even guide you to the source document where the information can be found. But what if you want information from across multiple documents? Or you want the information formatted in a particular way, like providing information in a tabular format?
Again, this is where generative AI changes things.
Microsoft Copilot for Microsoft 365 can work across Microsoft 365 applications—Microsoft Word, PowerPoint, Outlook, Excel, and others—to analyze, provide insight, and pull together information allowing you to access and manage all your content in one place.
While this approach allows you to look across documents you and your colleagues are using today, organizations are also seeking to unlock data in documents going back many years. Examples include understanding recipes and ingredients previously experimented with; attaining insight into previously run marketing programs or attaining perspectives on previous supplier negotiations in preparation for upcoming discussions. These are all use cases where the knowledge is spread across disparate locations and systems.
Already, several organizations have used generative AI to help improve the employee experience. Heineken, for example, has used Azure OpenAI Service and its built-in ChatGPT capabilities to build chatbots for employees, while also using other Azure AI Services to bring innovation to existing business processes.
Customer conversational experience
Solving a problem for your customer is a major way to differentiate your business from that of the competition.
A few years ago, when bots emerged, they offered the opportunity to allow a customer to get help without the need for a human. But the challenge was always that bots were limited by the topics and actions that your bot was configured for. In-short, they did not feel human enough.
Consumers often want help, advice, or inspiration with their purchases but without visiting a store this can be tricky. These ‘human-like’ interactions are so important that stores have invested heavily to save store associate time—freeing them to help customers.
Online this becomes difficult. But what if you could replicate a human expert who can help, advise, and inspire? One which could be available 24 hours a day to all your customers online?
This is where generative AI can power and dramatically enhance your Customer conversational experience.
In January 2024, Microsoft launched (in public preview) a copilot template on Azure OpenAI Service to build more individualized shopping experiences across existing web sites and applications. With this capability, retailers can build advisor type experiences for their customers who can engage in helpful and natural conversations and be guided to precisely the product they need. Help, advice, and inspiration all in one place.
Illustrating how this approach can differentiate, Carrefour launched their Hopla bot to help with what many consider a difficult domestic task—menu planning. After selecting the store where you want to do your shopping you can ask Hopla for a meal idea, based on your family size and budget. When you are happy with the suggestion the ingredients are displayed, considering assortment and availability at your chosen store. From there you can even add the products to your basket and transact for delivery or pick-up.
Carrefour built this using Azure OpenAI Service to access OpenAI’s GPT-4 technology. The solution respects confidentiality and compliance—leveraging Microsoft Azure data security, reliability, and confidentiality features, to ensure compliance with general data protection regulation (GDPR).3
Hopla is a great example of how AI can enhance customer experience and convenience, while also boosting sales and loyalty for retailers. By using OpenAI’s GPT-4 technology, Carrefour was able to create a bot that can generate natural and relevant meal suggestions based on user preferences and store availability.4
When they announced the launch, Carrefour said that customers will be “able to use this natural-language AI to help them with their daily shopping. They will find it on the site’s home page and will be able to ask it for help in choosing products for their basket, based on their budget, food constraints they may have or menu ideas.”3
This is a great example of how AI can help retailers differentiate themselves in a competitive market and offer personalized solutions that meet customer needs.
Generative AI provides more possibilities than can be addressed in a series of blogs. Understanding what others have done can help guide your thinking and approach. The level of creativity increases daily, and we will all watch the space with anticipation of the most impactful use cases for retail and consumer goods companies.
Transform your business with AI solutions from Microsoft
Microsoft AI solutions
Learn moreVisit the Microsoft Cloud for Retail website to learn more about how AI and generative AI capabilities are helping retailers and consumer goods organizations transform their businesses. Learn about Microsoft’s commitment to making sure AI systems are developed responsibly and in ways that warrant people’s trust.
1The economic potential of generative AI: The next productivity frontier, McKinsey.
2New technologies to take knowledge management in procurement to the next level, CPOstrategy.
