Introducing Glean Apps and APIs: Build custom generative AI agents securely at scale

When we envision the future of generative AI at work, many of us see everyday users and developers empowered to create domain-specific custom AI assistants, copilots, chatbots, and agents that automate everyday work, take action in multi-step workflows, and deliver proactive insights with AI across various applications. It’s the next step in the progression of generative AI for work. However, given the challenges of connecting AI agents with enterprise data and ensuring they deeply understand that data, deploying agents in enterprise environments is often prohibitively difficult. 

Today, Glean is launching the best way for businesses to quickly build and deploy secure, customized generative AI experiences grounded in their enterprise data, on one centralized AI platform. We’re excited to announce the general availability of Glean Apps, Glean Actions, and Glean APIs. These latest features will empower users to leverage the most advanced large language models to create their own custom generative AI apps and agents on top of the secure, permission-aware data connections and best-in-class retrieval engine of the Glean Platform. 

The complete platform for tailor-made AI experiences

The Glean Platform provides a suite of powerful tools that establishes a strong foundation for enterprises to develop AI experiences that address a wide spectrum of use cases. That includes everything from simple reference or chatbots that provide answers based on a limited corpus, to sophisticated agents capable of complex specialized tasks like delivering competitive intelligence solutions or debugging code. 

Let’s take a look at some of the key features in Glean’s AI platform:

Glean Assistant

Glean Assistant delivers an enterprise-ready digital workplace assistant powered by next-generation AI, featuring levels of accuracy and security not possible with general-purpose models on the market. Glean Assistant comes with a full understanding of all the content, context, and people within your company, strict adherence to permissions and data governance rules, along with full referenceability for all results generated.

It's a turnkey solution that's always grounded in your own enterprise information, scales with enterprise needs, and comes with over 100 pre-built connectors, ensuring easy implementation without the need of professional services, data training, or fine-tuning. 

Glean Apps

Glean Apps empower everyday users to create no-code custom generative AI agents and assistants tailored to specific business needs. Configured with natural language, Glean Apps can be constrained to reference a subset of company data and can take proactive action when triggered by specific events. These apps combine the comprehensive awareness of an enterprise assistant with the expertise of topic-specific agents.

Glean Apps can be customized in tone and format, can be easily accessed and set up without coding, and are published in the Glean App Library for easy discovery by other users in their organization. Additionally, Glean Apps can integrate into communication tools like Slack, enhancing real-time responses in specific channels.

Glean Actions

Glean Actions enable Glean Assistant and Glean Apps to directly take action on a user’s behalf within a company’s connected applications. Actions could involve reading data from an application, then executing a specific task within an integrated application, or could involve creating Jira tickets, publishing new content, or searching for code. 

Actions created in-house are also accessible to other collaborators within each organization through the Action Library. Users also have access to Glean’s Actions SDK, which enables them to develop custom actions tailored to their needs.

Glean APIs

Glean’s platform also offers robust APIs that enable users to leverage the enterprise-grade connectivity, retrieval, and analysis capabilities of the Glean Platform to build custom applications suited for their team’s specific needs. Users looking to incorporate generative AI without the complexity of developing a retrieval solution or chat function from scratch can use these turnkey solutions to embed Glean’s AI functionalities and foundations directly into the applications they use most. 

Glean currently offers two APIs for building generative AI experiences:  

• ‍Chat API enables developers to bring the power of generative AI experiences, grounded in company knowledge, directly inside other company applications. Beyond building tools like industry or workflow-specific mobile apps with turnkey generative AI solutions, developers can also build custom UX around those experiences, embedding Glean into any surface like intranets or customer service portals. 
• ‍Search API provides the power of Glean’s best-in-class search to power enterprise-wide retrieval augmented generation (RAG) in custom generative AI applications. Have your own AI stack but lacking a powerful and secure enough retrieval solution to enable RAG for all enterprise data and activity? Use Glean’s Search API.

We’re also excited to announce today out-of-the-box support for Claude 3 Sonnet on Bedrock and Gemini 1.5 Pro on Vertex AI, joining the existing support for OpenAI’s GPT 3.5 and 4 models. With the rapid advancement of large language model (LLM) capabilities and improvements, we're committed to ensuring that our customers benefit from the best across various LLMs by taking a model-agnostic approach. Learn more about our approach to multi-LLM support here.

An enterprise-ready foundation for the future of AI

Integrating agents and custom AI experiences into enterprise environments is easier said than done. Building agents from scratch in-house requires enormous financial and resource investments, particularly when it comes to curating the retrieval augmented generation (RAG) corpus for each separate agent. Buying agents individually is a costly strategy as well, as they’ll need to be purchased separately according to their function and use case. Most third-party agents also won’t have a deep enough understanding of all the people, data, and activity within your organization to generate accurate and useful results. 

Glean is uniquely positioned to solve for these issues. We connect to all of your data sources securely, without restrictions to a particular vendor’s ecosystem. Glean also handles all the enterprise technicalities and requirements around maintaining permissions and keeping data up to date, so you don’t have to. Everyday users and developers can now directly leverage the power of Glean’s retrieval and indexing systems to build domain-specific solutions of their own—all without compromising on privacy, security, speed, or cost. 

During the Glean Apps and Glean APIs preview, over 100 Glean customers – including leading technology, media, telecommunications, manufacturing, industrials, and professional services companies – created and deployed 1,000+ unique AI apps and agents. The Glean Platform is now in use by hundreds of customers, including BILL, Databricks, Duolingo, Grammarly, and Sony Electronics. I can’t wait to see what others build.

Looking to get started with Glean? Sign up for a demo! Already a Glean user? Get a closer look at how to use Glean Apps and Actions today.