Sell More With ChatGPT Product Recommendations

Introduction

OpenAI recently announced that ChatGPT will be adding product recommendations as part of its search experience. This will act as a personalized, conversational shopping assistant and display relevant product options in visually rich carousels, provide additional product details, and link users to websites where they can learn more or make a purchase. A 2024 study by Bain showed that eighty percent of consumers now resolve 40% of their online search queries without clicking any links. The same study also found that 42% of people who regularly us generative AI tools also use them for shopping recommendations (WSJ). With this in mind, being visible on these platforms is crucial for online retailers.

How It Works

There are three main inputs ChatGPT will consider when providing these product recommendations.

1. Structured Metadata

2. Previous Model Responses

3. OpenAI Safety Standards

Structured metadata is the one that retailers will have the most control over. This includes product specifications such as the listed stack height, drop, distance suitability, etc. Additionally, ChatGPT can also make inferences from reviews or published content. For example, if a user is asking for a budget-friendly shoe recommendation, ChatGPT could recommend one where an online reviewer mentioned “good value for money” instead of the cheapest option.

Practical Example

To illustrate how this could play out, imagine the following scenario.

A runner asks ChatGPT to help them create a four month marathon training plan with the goal of running under 3:30 at the Toronto Waterfront Marathon in October. When creating this training plan, they mention that their current PR is 3:32 from the year prior and they like to train in the New Balance 880v13. ChatGPT knows:

• This is a serious runner

• Better gear could provide the marginal gains necessary to help them reach their goal

• They have some level of brand loyalty to New Balance

• They will be in Toronto

Closer to the marathon, this runner asks for racing shoe suggestions. Pairing these considerations with online product reviews and descriptions, ChatGPT could recommend specialty retailers around Toronot that stock the New Balance Supercomp Elite V4, or another similar shoe.

Why This Matters To Retailers

By providing organic, conversational recommendations, AI is influencing buying decisions before users visit retailer sites. Once a consumers has signalled purchase intent, retailers can have their products displayed before them.

In essence, AI is acting as an extension of a retailers salesforce if they optimize for it correctly.

How Retailers Can Take Advantage Of This Today

1. Optimize Metadata

Ensure your product pages include detailed specs and use customer-centric language. You can also leverage customer reviews to highlight value or niche use-cases.

2. Ensure your site is crawlable by OpenAI

ChatGPT uses a web crawler called OAI-SearchBot to find, access, and surface information in ChatGPT search. This is managed in your site’s robots.txt file; speak to your webmaster to ensure the OAI-SearchBot has access.

3. Sign up to be notified when product feed submissions open

Similar to Google Shopping listings, ChatGPT will eventually allow merchants to upload their product feeds in order to provide more accurate, up-to-date listings. Sign up by filling out the form here: https://openai.com/chatgpt/search-product-discovery/

Additional Strategic Considerations

Cross Selling and Bundling

Consider listing bundles (ex. A full race day kit) or including metadata for related products. This could potentially boost cart value by encouraging ChatGPT to surface larger product packages where relevant.

Local vs. Online Retailer Advantage

Local retailers can use local inventory to their advantage. Highlight services such as local pickup or try-on options to compete with larger eCommerce players. 

Content Marketing and AIO (AI-Optimized Content)

AIO is becoming an increasingly important consideration in every brand’s content strategy. Like SEO, you need to start with a strong content strategy that provides value to your readers. In the case of running brands, resources like buyer guides, race day checklists, and expert reviews aren’t going away. What marketers need to be careful about is ensuring that these resources use clear, structured language that AI can easily parse. This will help ensure that AI is providing users with the information you want it to. Remember: garbage in, garbage out.

Update: Here is a great article outlining how to get started with AIO in 48 hours: https://www.samhogan.sh/blog/a-new-form-of-seo