Glean Technologies Inc., the creator of a powerful generative artificial intelligence-powered enterprise search application, is expanding into “agentic AI” territory with the launch of its new Glean Agents platform.

Unlike other agentic AI platforms, it allows enterprise workers to create what Glean calls “horizontal” AI agents that can access broad data from across the entire organization, as well as public data from the internet. This means they have the knowledge required to act “horizontally,” across every department within the company.

Glean is best known for its conversational enterprise search platform, which is a kind of ChatGPT for work, boasting features that are unique to business scenarios. For instance, it supports multistep prompts that allow it to automate complex workflows and it has connectors so it can pull data from more than 100 enterprise applications.

The company is using those connectors as part of its new Universal Knowledge service, which underpins the Glean Agents platform by bringing together organization-wide data with information from the internet. This means its AI agents can access all of the information required to perform a wide range of complex business tasks.

AI agents are a more advanced kind of AI application that go beyond what traditional chatbots can do, responding to questions and creating images. They’re designed to perform various actions on behalf of users with minimal supervision, such as automating responses to customer’s emails, qualifying sales leads, providing technical support to users, booking hotel rooms, scheduling appointments and so on.

Glean says its horizontal AI agents provide an advantage, because most existing ones are limited to isolated departmental applications. By giving them access to a broader set of data, they’ll be able to scale horizontally to perform tasks across every department, boosting their potential and increasing productivity.

The Glean Agents platform consists of a natural language agent builder, which allows employees without coding skills to create AI agents simply by describing what they’d like it to do. The platform has the ability to understand the end goal, and then it automatically creates the agent, using company knowledge to facilitate the multistep workflows required to achieve it. Users can select from a catalog of large language models to power their agents, and then refine the parameters to get it just right.

In addition to building their own agents from scratch, users can select from various prebuilt agents and fine-tune them to their needs.

Each AI agent benefits from Glean’s LLM-agnostic Agentic Reasoning Engine, which enhances the capabilities of the underlying LLM with advanced reasoning functionality. This can help them to perform more sophisticated automations and make smarter decisions by understanding both context and the desired outcomes, using a library of actions to execute different tasks.

What really sets Glean’s AI agents apart is the Universal Knowledge service, which gives them the ability to analyze structured data from databases and applications such as Databricks, Salesforce and Atlassian, and combine this with unstructured data such as an organization’s documents, and information from the web. Moreover, they’re not just limited to “old” data, as they can search both internal systems and the web in real time to find the most up-to-date information whenever it’s required.

Andy Thurai, vice president and principal analyst at Constellation Research Inc., told SiliconANGLE that the Universal Knowledge service can help to solve a key painpoint in many LLMs, namely their inability to access enterprise data.

“Most LLMs in use today have their knowledge cut off, which limits them to specific information up to a certain timeframe, but not beyond it,” Thurai pointed out. “AI agents can improve on this by enabling users to search for the latest corporate information that’s not embedded in the LLM itself.

The analyst said Glean stands out for its abililty to connect its AI agents to more than 100 popular enterprise applications and data stores, as this means they should be able to answer some very specific questions and execute tasks that require knowledge of the company’s private business data. This, in addition to its no-code agent building tools, can help Glean to stand out from the crowd, he said.

“While Glean is not alone in the crowded agentic AI segment, it is offering some unique features that make it a compelling option for enterprises that need to build reliable AI agents quickly,” he added.

Glean also provides assurances on security, citing its active data and AI governance capabilities, which implement continuous scanning to identify any potentially sensitive information and flag it for administrators, who can then check to see if it should be sharing that info.

Glean founder and Chief Executive Arvind Jain said his company’s vision is to empower every worker with the ability to create domain-specific AI assistants and agents that automate their everyday work.

“They take action in multistep workflows and deliver proactive insights across all applications,” he said. “This approach ensures that our AI agents are not only practical, but also deeply integrated with enterprise data to provide meaningful and secure solutions to complex workplace challenges.”

Image: SiliconANGLE/Dreamina