edison lightbulb

Coveo Launches Self-Learning Search for the Enterprise

3 minute read
Erika Morphy avatar
Its new product Reveal is based on machine learning plus models developed from data gathered via its analytics offering.

Coveo, an enterprise search provider headquartered in Quebec City, Canada, rolled out a cloud-based search product yesterday that's based on machine learning technology. It builds on Coveo's strong credentials in this space, namely its advances in indexing across multiple, disparate sources of information and its insights about user behavior and analytics.

The product is called Reveal, an apt name given Coveo CEO Louis Tetu's approach to search. "For us, search is how to surface the most important and relevant piece of information for the user, no matter where it resides," he told CMSWire.

Reveal overlays machine learning on top of Coveo's behavioral click learning analytics, a product it launched two years ago that has been deployed by a number of marquee corporate names including Salesforce (a strategic partner of Coveo), Sitecore (ditto) and GoPro, among others.

'Massive Data Points'

Companies typically deploy Coveo as the backbone of an internal customer relation operation or to help manage a community of users. Here is where Coveo's indexing and search expertise is on full display: it can cull data across a wide variety of channels — from YouTube to service manuals to internal databases — and surface it to the agent.

Two years of Coveo Analytics in operation have also garnered a rich data store about user behavior and insights to be used as models for Reveal. 

coveo revel screenshot

"What we have accrued are massive data points around everything a user does," Tetu said.

The self-learning or machine-learning component is how Reveal fine-tunes the delivery of relevant information. It starts matching results with relevance upon the platform's launch and users can see improved results within a matter of weeks, Tetu said.

Three beta companies, one of which was VMware, reported that they experienced:

  • improvement in click throughs of 10 percent to 20 percent
  • improvement in click ranks of 5 percent to 15 percent

"Information from Coveo Usage Analytics has been extremely valuable to us, helping to increase the usability and self-service capability of our partner community," said David Wylie, manager and central partner operations at VMware. "Once we've deployed Coveo Reveal completely, it will ingest the analytics on search behavior and use it to automatically tune results."

Learning Opportunities

Preliminary results look very good, Wylie added.

$75M in Funding

Earlier this month Coveo secured $35 million in Series D financing, a round led by IQ Venture Capital with participation from existing investors, including Tandem Expansion Fund, BDC Capital, Propulsion Ventures SEC, Fonds de solidarité FTQ as well as CEO Tetu.

Coveo has raised $75 million in financing to date.

Tetu said the company will be using the proceeds to develop new use cases to launch into new lines of business.

"I won't get into specifics, but I will say at a high level Coveo wants to be active in two areas. One is customer engagement, by integrating customer experience from the front end to customer services communities and to the CRM system," he said.

"The second is the digital workspace. We can not only reach the right information for workers but also recommend to them which is the most relevant information."

Title image by Josh Byers