Join top executives in San Francisco on July 11-12, to hear how leaders are integrating and optimizing AI investments for success. Learn More
Presented by Glean
Generative AI can unlock the full potential of data in enterprise environments for the employees who rely on it. In this VB Spotlight event, learn how generative AI has transformed enterprise search, improving productivity, building better business outcomes, and more.
Enterprise search is a growing pain point. The explosion of SaaS tools over the past decade has brought a sophisticated array of solutions that have changed how work is done – but has also brought along data fragmentation. Employees are working in multiple, disparate applications, creating content in one, communicating about it in several others, looking for background information in yet another, and so on. No one is clear where documents live, where information can be dug up, whether it lives in someone’s head or is hidden somewhere on the network.
Join us in San Francisco on July 11-12, where top executives will share how they have integrated and optimized AI investments for success and avoided common pitfalls.
“The pain points come from not knowing how to navigate this sprawl of software and information, and the mental overhead required to remember these things cuts across functions,” says Eddie Zhou, founding engineer, intelligence, at Glean. “It makes onboarding new hires a massive pain point as well, especially in a hybrid or remote environment — it’s hard to know what you should be looking at and where you should be looking for it.”
Generative AI, which has been blowing up the headlines, has made it possible for users to interface with an enterprise work assistant in a much more natural way, taking a large chunk of cognitive overhead away. It makes enterprise search feel like the web searches they’re used to, lowering the barrier to knowledge.
The evolution of enterprise search solution
Companies have been trying to tackle the challenge of enterprise search for decades, mostly with custom internal tools, but the technology to create a comprehensive solution hasn’t existed before now. A general standardization of tools across organizations – most companies use the Microsoft suite, for example, or Jira, etc. – was a step toward scalability. Artificial intelligence was another step forward, but the main challenge in enterprise search is sparsity: a much smaller set of documents from which to train a model.
“The advent of large foundation language models in 2018 made it possible to bring knowledge from the web and from larger sets of data to make enterprise search, which operates on a much smaller set of data, work closer to the way that people have come to expect,” Zhou says.
Building the kind of system that learns and works out of the gate was a key turning point, and many enterprise search players today are operating on that model, building manual heuristic systems that need to be plugged in and hand-tuned. And now, generative AI is a leap forward, bringing a new kind of intelligence to a plug-and-play search engine.
How generative AI transforms enterprise search
Conversational AI essentially peaked in 2016 and 2017 and then seemed to peter out, because many promises were made about the potential of the technology, before the technology was actually sophisticated enough to keep them. Today, with ChatGPT going mainstream, the technology is significantly more advanced, and the vision of a conversational agent in the work setting is a much more real possibility, Zhou says.
It’s about giving people access to information they need in a way that feels intuitive. And it can bring users the information they need, when they need it, comprehensively searching apps across the company, understanding context, language, behavior and relationships to find personalized answers. It can surface knowledge and even connect users to the people who can help answer questions or accomplish tasks.
A solution like Glean connects to all of an organization’s data sources, crawling the content and indexing all metadata that exists for those data sources, such as links between documents and messages, authors, access permissions, activity surrounding content, by whom, from where and when. For instance, while Slack search is useful to surface an old message, that search can’t follow a link to a Google Drive and index any of the information in those documents. Being able to connect to everything that a given company might have knowledge in, makes the search engine’s knowledge complete. Leveraging data from multiple sources means the engine is always learning and makes the search stack better.
“That completeness really is necessary to deliver a search experience that works,” says Zhou. “When a given employee comes to their keyboard, in their mind they have a mental model of all the ways their data is connected. The system that they’re working with also needs to have that.”
Securing data with a trusted knowledge model
The conversation around trust and ethics in generative AI is crucial, Zhou adds, and the trusted knowledge model is fundamental to delivering a generative experience in the enterprise.
It’s built into how the platform indexes information. For each data source it connects to and each document it crawls, it also natively crawls its layers of permissions. This unified view of who a user is across data sources means a search will only turn up the documents and information they have access to. Referenceability, or transparency into where the generative model found that information, means a user can trust the answers they receive.
“For us, the foundation of the trusted knowledge model is permissions and data governance, and it’s fundamental to delivering a good generative experience,” he says. “Building on top of permissioned search also lets us ensure that we are providing relevant information, because we’ve understood who a user is, understood the language of a given company, plus the relationships between information and those people. Ultimately we are able to deliver a better end-to-end experience.”
To learn more about how generative AI unlocks the full potential of enterprise data, a closer look at the trusted knowledge model for generative AI, and more, don’t miss this VB Spotlight.
- Understanding the present and the future of AI in enterprise search
- Unlocking the full potential of data in enterprise environments with generative AI
- Recognizing the importance of a trusted knowledge model for generative AI
- Facilitating information access and discovery to improve employee productivity
- Creating more intelligent, personalized, and effective experiences
- Phu Nguyen, Head of Digital Workplace, Pure Storage
- Jean-Claude Monney, Digital Workplace, Technology and Knowledge Management Advisor
- Eddie Zhou, Founding Engineer, Intelligence, Glean
- Art Cole, Moderator, VentureBeat
VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.