AI in e‑commerce

According to a Salesforce study, revenue increases by 26% when a store is personalized and uses AI.

AI will help e-commerce businesses by saving them time on tasks that are often time-consuming but essential for an e-commerce site.

Examples of 5 use cases from among those mentioned above

Searchandising and Merchandising

AI has been used on the site for many years to offer more personalized shopping experiences tailored to visitors’ needs (whether they are anonymous users or customers), to improve relevance, reduce the number of clicks needed to find the right product, and thereby increase the rate at which items are added to the cart.

In the search engine feature,

• Generative AI enables the generation of synonyms in the AI-powered search engine
• Semantic AI is used to understand what a user is searching for in the search bar and to make search results more relevant

E-commerce and marketplace solution providers are now integrating tools such as Adobe Commerce Cloud with Live Search, Intershop following its acquisition of the Sparque solution, Mirakl with Target2Sell, and, of course, specialized solutions like Algolia, Sensefuel, EarlyBrids/Attraqt, and Nosto…

At SoBrico, Algolia has been used for the past 2–3 years for searchandising a product catalog containing more than 100,000 items. The goal was to reduce the number of clicks needed to find the right product. Today, 70% of traffic comes from the search engine, which accounts for 50% of revenue.

Content & SEO

For content creation and SEO optimization, AI can help you develop a sitemap, write meta tags, create effective internal linking, and generate content ideas for product pages as well as for how-to and brand content pages.

The ChatGPT AI Content Generator module, developed by Amasty and validated by Adobe Commerce Cloud, enables, for example:

• Creating and generating content for product and category pages
• Generating Structured Data to Optimize SEO

This is an important point to keep in mind because, for SEO, it’s crucial to stand out in order to improve your search rankings. Search engine algorithms are based on originality, relevance, and the “EEAT” criteria: Expertise, Experience, Authority, Trust. Be careful not to fall into a “Full AI” trap (hallucinations, identical syntax structures, etc.)

• Including personal stories and anecdotes
• Adopting the brand’s tone
• Drawing parallels with recent contexts/events
• Citing verified sources

Product Recommendations

SoBrico recently decided to use the product recommendation feature offered by Algolia because this feature has the advantage of taking into account both the user’s search behavior and the recency and freshness of the site’s content. To make this possible, it was necessary to implement the Events API provided by Algolia to track all events related to navigation, purchases, content, products, and customers.

So Brico decided to use this feature in the form of product recommendations on the home page, product pages, and the “Add to Cart” page.

Chatbot

Often criticized for its customer experience—which relied on a simple decision tree—the arrival of AI has clearly changed the game. The “Hello Casto” module is proof of this. By implementing an intelligent chatbot, Castorama sought to offer a new and innovative experience, ranging from personalized advice to product recommendations.

The goal was to address an online “pain point” by offering a genuine value proposition: to try to replicate online the experience of talking with a sales associate in a store.

When you test it, the questions are intentionally open-ended, which required a high level of ability to gather information, put it into context, and engage in a conversational exchange.

It took 6 weeks to put the 1stversion version online, which involved integrating innovations and service-oriented AI, devising use cases based on user needs, and designing the digital experience and interface for this new service.

This service was tested with about 20 users and testers before its launch in the power tools product category. It was also initially populated with Castorama’s brand voice, internal knowledge, and informational guides created for sales associates, while also anticipating non-standard use cases.

Technically, Castorama leveraged the Kingfisher Group’s framework called Athena, which utilizes three of the leading language models on the market, enabling the system to deploy the most relevant LLM based on specific use cases and ensuring regulatory and editorial compliance in communications: Google’s Gemini, OpenAI’s ChatGPT, and Anthropic’s Claude.

In terms of customer usage, Castorama realizes that this will take time. Even though the site receives a high volume of traffic (10 to 15 million visits), current usage patterns mean that users today aren’t accustomed to interacting with an avatar. They use traditional navigation and the search bar.

Inventory Optimization

We’ve just looked at use cases above that have an impact solely on users. But AI also provides operational benefits with higher added value for e-commerce businesses in the area of inventory management. Through machine learning algorithms and the real-time analysis of massive amounts of data, AI has brought a proactive and dynamic approach compared to the traditional methods used until now (manual processes and statistical rules).

By leveraging AI, companies are now optimizing their inventory levels based on sales data and third-party data (weather, seasonality, economic conditions, and market trends). This allows them to reduce both excess inventory and the risk of stockouts, thereby lowering storage and carrying costs.

AI also helps them optimize their warehouse operations. This improves the efficiency of storage and retrieval processes, reduces labor costs, and increases productivity.

In conclusion

Generative AI is undeniably a tool that streamlines work, delivering unmatched efficiency and productivity. It can produce analyses, gather information, and generate content in impressive ways. But human expertise cannot be displaced; it is irreplaceable.

The limitations of AI are evident, particularly in its limited learning capabilities, its “hallucinations,” and its difficulties in interpreting, selecting, and processing reliable sources of information.

3 Things to Keep in Mind When Adopting AI in Your Business

• Implement a change management process within your organization to explain what AI is, how it will be used, and how the roles and responsibilities of the people who will be using it will change
• Ensure the security, hosting, and architecture of your data and the AI solutions you use, as well as the associated intellectual property rights
• Don’t fall too far behind—which doesn’t mean “going too fast.” It’s important not to rush, but to think carefully about your goals and how to implement AI.

AI is a tool, not a magic solution

Keep the customer at the center of your strategy and people at the heart of this transformation.

The use of AI must be guided by ethical principles

Maintaining control over your content, data, and user journeys is a prerequisite.

Don’t fall behind

Mastering the customer experience is a major competitive advantage that sets you apart from your competitors

Need advice on integrating the AI
into your business?

A project? Any ideas?

Talk to our experts!