Coveo is infusing SAP Commerce Cloud search with AI to bring targeted recommendations to customers.
Coveo Solutions, based in Quebec City, provides AI search that sits on top of enterprise platforms to deliver personalized results for B2B and B2C e-commerce. Its Search and Recommendations engine is built on the Coveo Relevance Cloud platform, which uses machine learning and deep learning to index secure data across SAP and other ERP systems, knowledge bases, CRM, and other enterprise systems.
For SAP, the Coveo Search and Recommendations engine indexes SAP Commerce Cloud catalogs and layers in AI with the aim of providing a personalized experience, according to the company. It was recently certified as an SAP Endorsed App and is available to customers on the SAP Store.
The AI-infused search works similarly to a streaming service such as Netflix that adjusts recommendations based on individual user behavior as well as collective behavior from across the user base, said Louis Têtu, CEO of Coveo.
"AI learns from every interaction to create your next one," Têtu said. "It learns from the sum total of interactions and outcomes of all the consumers to, within 10 milliseconds, dynamically reassemble the content, the navigation, the recommendations and the ranking of products for the consumer."
Chamberlain Group, a Chicago-based company that makes garage door openers and smart building access products, integrated Coveo's search functionality into its SAP Commerce Cloud B2C and B2B e-commerce sites.
Chris KingSenior manager for digital experience and media, Chamberlain Group
The company had used Coveo search for several years on its customer support site, and in 2021 wanted to improve the search and recommendations capabilities in the consumer e-commerce site, according to Chris King, senior manager for digital experience and media at Chamberlain.
Chamberlain was using SAP Commerce Cloud's built-in search functionality, but had to heavily customize the search results. Eventually, the company's requirements outstripped the technology's ability to provide the customization it needed, King said.
"We wanted to be able to leverage machine learning so that the platform was taking in consumer and customer search behavior and improving results over time as more and more data was collected," he said. "[We also wanted to have] more granular control around how we merchandise the ranking of our products in search results and in our product listing pages."
In the year since Chamberlain implemented Coveo's AI search on SAP Commerce Cloud, the company has seen higher conversion rates with consumers and reduced search queries that don't return results, King said.
"The other thing, which is particularly pronounced with the B2B side of the business, is that we use surveys to get at some of the qualitative analysis on things," he said. "We've seen a significant drop in issues where someone conducted a search and wasn't able to find what they were looking for."
The Coveo approach differs from generative AI such as ChatGPT because it brings relevant data into the process that's able to better target what a customer is looking for, according to Têtu.
Generative AI platforms can be useful for tasks such as generating recipes, as with the integration of ChatGPT into Instacart, but enterprises need relevant and trusted recommendations, he said.
"What's important is that for most enterprises, veracity, factuality and security are critically important," Têtu said. "ChatGPT 'hallucinates,' it lies in a very elegant way like Shakespeare would be able to lie, but it's factually often incorrect."
E-commerce sites need to stand out
Coveo's AI-based approach can be a significant tool for e-commerce sites that need to compete with Google for search or Amazon for business, said Brian Kilcourse, managing partner at Retail Systems Research, a retail technology consultancy.
E-commerce sites need to make their products stand out and have relevance for consumers, but consumers want to shop without providing too much personal information, he said. Coveo's AI modeling can identify consumer intent and get to what they want.
"AI is really good at doing this. It can identify -- based on a few behaviors that you've engaged in online -- that you're who you are and you're likely to be looking for this kind of a product," Kilcourse said. "They can appear to be very personalized even when you haven't identified yourself, which it does without being creepy."
All e-commerce providers are trying to add this kind of search and recommendation capability, and enterprise vendors are working on the technology, he said, but Coveo has an advantage of identifying the issue early and being devoted to it.
"A broad offering like SAP or Salesforce has more features and functions that they have to be concerned about, whereas Coveo is razor-focused on this issue," Kilcourse said.
Jim O'Donnell is a senior news writer who covers ERP and other enterprise applications for TechTarget Editorial.