Feature
Ask knowledge workers about their experiences with enterprise search, and the feedback is often grim. Despite all of the advances made since FAST Search, Autonomy, IBM’s STAIRS (Storage and Information Retrieval System) and others were first introduced, finding accurate information in the right context at the right time in real time remains difficult within most organizations.
Or as Deep Analysis founder Alan Pelz-Sharpe put it: “Federated enterprise search fell flat on its face."
This may be changing. More on this later.
Many of the early vendors in the enterprise search business have since got out or been acquired. Take Google Search Appliance, which announced its exit in 2016. While it might seem that googling for information at work should be as easy and straightforward as it is in your private life, Google results are based on popularity. "Enterprise search has to be based on what’s accurate, what’s the latest, not on popularity ...,” said Pelz-Sharpe. That includes ensuring data accuracy, protecting data privacy, controlling access to information, following legal and regulatory requirements and more. The goal is to deliver and keep information secure, compliant and accessible to the right people.
What Happens When Information Can't Be Found
The consequences of poor enterprise search can be costly. Take the Boeing Dreamliner construction project, for example. The inability to quickly locate and share information across teams and locations slowed down its development process. The project accrued cost overruns due to the need for rework and missing or inaccurate information. Quality control was an issue because of difficulties with tracking changes and teams working on different versions of documents.
The enterprise search platforms of yesteryear could have helped mitigate some of these issues. By connecting disparate data sources, technologies like Elasticsearch, OpenText IDOL and Sinequa among others might have provided a unified view of project data, helping ensure Boeing workers were all on the same page. Their ability to understand complex relationships within data might have enabled proactive risk identification while data cleansing could have enhanced data quality, minimizing errors and delays. But all of this depended on having a strong enterprise search team managing the platform, a key element too often downplayed in vendors' sales pitch.
Cognitive Search vs. Cognitive AI
Over the years, other solutions tried to solve the root issue of finding information. Cognitive search (also known as insight engines) made its mark in 2017 when Forrester published its first wave report about the AI-flavored enterprise search platforms. Cognitive AI's origins in the context of enterprise search are murkier.
So what’s the difference between cognitive search and cognitive AI?
“While Cognitive Search is a step above traditional search, using AI to deliver more accurate and contextual results, Cognitive AI takes this capability to the next level,” Keri Rich, vice president of product at Lucidworks told Reworked. The latter adds human-like reasoning and natural language conversations to search, shifting the experience from basic queries to dynamic interactions. "Think of Cognitive Search as your librarian, while Cognitive AI is more like having a consultant who can find information and help you reason through complex problems," said Rich.
Large language models (LLMs) have now been, or will soon be, incorporated or added into many cognitive search and enterprise search solutions. Theoretically, this incorporation could have aided the Boeing 787 Dreamliner project by facilitating seamless communication between Boeing and its global suppliers with real-time language translation, and by analyzing vast amounts of content, including technical documents, supplier performance reports and market trends to identify potential risks and proactively predict delays. Boeing could have made more informed decisions as a result regarding production schedules, resource allocation and risk.
The aforementioned providers of cognitive search platforms have begun to leverage generative AI technologies to “grow their businesses as well as address some of the more complex enterprise challenges behind a frontend experience — such as document preparation and processing, privacy and access controls, understanding of user intent, and data connectors to enterprise systems,” Rowan Curran, senior analyst at Forrester Research, told Reworked.
Lost in the Clouds
While most companies don’t undertake projects as large as Boeing’s, enterprise search can still benefits workers and teams working at a smaller scale. Take this for example: You’re a top-tier account executive who can't find an important sales proposal created by your team. Your prospective customers, who arrived early, are waiting for you in the conference room down the hall. You frantically search Slack, through email attachments and all four of your company's cloud services. Nothing. Just as you’re about to call your boss to see if they have a copy, a hit pops up on your screen, and soon after, another, and another one after that. They are all named something different and seem to have been saved at around the same time. How could this be happening?
It's problems like this that Glean, an AI-powered enterprise search platform, was built to solve. Its AI assistant and Work AI platform make it “easy for workers to find answers, generate content and automate work,” Paloma Ochi, head of product marketing at Glean told Reworked.
Unlike some cognitive search tools that were first built for outward-facing sales and marketing managers, Glean was built specifically for enterprise searches. "We do it in a way that is safe, secure and permission aware,” said Ochi. Glean leverages a knowledge graph that captures all content, people and activity within an enterprise and understands how content is being used, by whom and what they are doing with it. This helps it deliver precisely the results that the information seeker requested and the sources for the data. The average answer is extracted from three or more data sources.
Beyond Finding Information
Some cognitive/enterprise search platforms work together with generative AI tools to create insight-filled content. While cognitive AI quickly processes large volumes of data, generative AI creates summaries, reports or presentations based on that information. Glean, for example, assists with content creation, authoring and publishing blog posts, producing articles, correspondence, meeting briefs and more.
Several knowledge management platform providers offer enterprise search as a component. Bloomfire is one such example. The AI-powered platform centralizes enterprise information and makes it searchable via its deep-indexing architecture. The promise is that a worker can ask it a question, and it will both come back with an answer and explain how it arrived at it. “Bloomfire aims to get the right information to the right people at the point of inflection,” Dan Stradtman, CMO, Bloomfire told Reworked.
What’s somewhat unique about Bloomfire is that its AI encourages knowledge-sharing by making it easy to contribute expertise and encourage engagement. “When employees see their insights valued, they’re more likely to keep sharing,” said Stradtman.
What’s Next for Search?
Generative answering is still a fairly new term. Curran characterized it as “results being summarized as a narrative, detailed contextual list, or other format that is more specific and useful than a list of search results.” He explained that it has already helped significantly reduce the number of customer service requests that need to be directed to human agents and improved customer satisfaction. “It’s also supporting predictive maintenance applications by using an alert to trigger a search which then uses a language model to synthesize technical documents into a single procedure to resolve the raised issue,” he added.
What’s the future of cognitive and enterprise search? In an AI-driven world, we wouldn't hazard a guess. If you really want to know, ask AI.
Conference
Sep
16
HR Tech Conference Las Vegas 2025
Register
Conference
Oct
27
Gartner HR Symposium/Xpo Orlando 2025
Register
Read more about the world of search and knowledge management:
- Diagnosing Enterprise Search Failures — How to diagnose and improve search without changing the search engine.
- Knowledge Management Means More Than Just Mining Digital Exhaust — If we want machine intelligence to play a role in knowledge creation, we need conscious design, not lazy accidents.
- GenAI Can Improve Enterprise Search, But Remains a Work in Progress — Generative AI could overcome traditional problems associated with enterprise search. However, it still has a way to go before its fully functional.
