Crafting intelligent search within the enterprise at KMWorld 2024

Despite the dramatic introduction of AI to deliver information retrieval relevance, enterprise search remains a silent and significant drain on productivity and revenue.

At KMWorld 2024 in the Enterprise Search & Discovery track, Marianne Sweeny, principal consultant, Daedalus Information Systems and Bonnie Chase, director, AI product marketing, Vespa.ai discussed, “Intelligent Search Within the Enterprise.”

Sweeny explained how to create a road map for success, build a user-focused experience, and drive continuous improvement.

“What we need to understand is that information isn’t something that you can teach anyone, it’s an emotional construct,” Sweeny said.

Neuro intelligence and Neuro IR is feeding into the search landscape, Sweeny explained.

The user experience demands something new and different. If someone doesn’t find what they are looking for on your website, they will leave, she said.

Soft system methodology emphasizes looking at the whole instead of just a sum of the parts. The ultimate feedback loop is the person looking at the results.

“We need a way to evaluate those results and bake in the ways to make that happen,” Sweeny said.

Connect the dots with:

• Relationships (ontology)
• Organization (taxonomy)
• Naming (labels)
• Markup (schema)
• Equity data (validation)

“Following false prophets are only going to give you misery in the end,” Sweeny said.

Make sure to input risk mitigation, she said. The key final takeaways when perfecting search are to broaden scope of awareness, understand the landscape and influences, and embrace new tools and methodologies.

Chase noted that the rise of GenAI has significantly transformed the search landscape and believes that the future of search is multimodal. Limitations encountered today, particularly with PDFs, have caused knowledge managers to change strategies. But search should be tailored for future capabilities.

“This is tomorrow, this is 6 months from now, this is what you’re going to see next year,” Chase said.

New advancements, such as ColPali, will transform the world of information retrieval, she noted.

From a KM and search perspective, combining images and text leads to improved user comprehension and experience, ultimately improving self-service success. However, the tech we use today has some limitations.

ColPali is a method for retrieval and builds on two concepts: conceptualized embeddings from VLM and late interaction.

“Not only are knowledge managers able to create multimodal content, we can use that content information for retrieval,” she said. “It improves the accuracy and intelligence of your results.”

KMWorld returned to the J.W. Marriott in Washington D.C. on November 19-21, with pre-conference workshops held on November 18.

KMWorld 2024 is a part of a unique program of five co-located conferences, which also includes Enterprise Search & Discovery, Enterprise AI World, Taxonomy Boot Camp, and Text Analytics Forum.