Mattermost brings generative AI power to open-source collaboration

Collaboration platform vendor Mattermost has expanded its namesake open-source technology to  enable security conscious organizations to better integrate generative artificial intelligence (AI).

Mattermost’s open source collaboration technology allows organizations to communicate and manage workflows. As a service, the company provides cloud offerings that provide enhanced enterprise capabilities for authentication and performance.

Mattermost is vying for a portion of the expanding collaboration tools market, facing competition from major players such as Slack, Atlassian and Asana. As with nearly every other segment of IT, the collaboration tools market is now embracing the power of generative AI, and Mattermost is no exception.

The new generative AI expansion at Mattermost will enable better integration with OpenAI to allow organizations to build chatbots and gain insight into workflows. Mattermost is also providing what it calls ‘functional intelligence’, by way of integrations with ServiceNow‘s virtual agent platform, as well as technology from Ask Sage.

“What our customers care about is running secure collaboration in strict security environments, and they’re looking to rapidly bring AI into these environments,” Ian Tien, Mattermost CEO and cofounder told VentureBeat. 

Generative AI, functional intelligence and domain knowledge

There is more than one way AI is impacting the Mattermost platform for enterprise and government customers.

The first category is generative AI, where Mattermost now has a framework allowing organizations to plug-in different providers including OpenAI’s ChatGPT, as well private cloud large language models (LLMs). Generative AI can help organizations to summarize or create new content and insights.

Another important AI category that Mattermost now integrates is functional intelligence. Tien said that Mattermost’s integration with ServiceNow and its virtual agent is a prime example of this, providing the ability to augment employee and customer service.

The third category for Mattermost’s integrations is domain specific AI, where AI is deployed within a specific vertical domain or organization for privacy and security compliance. That’s where the integration with Ask Sage fits in. Ask Sage provides teams with fast data analysis and summarization capabilities to extract vital insights as situations change.

One example of how the new AI integrations will aid collaboration and workflow is something as basic as software installation, said Tien. For example, a system administrator looking to install a piece of software can now use AI to get answers and insights on how to best configure and deploy the software. The information the AI generates can come from public as well as private sources, delivered in a secure way to optimize the overall user experience.

Why data is the key to fine tuning AI

The market overall is shifting from an era where AI was model-specific to one where AI is data-centric, said Tien.

As such, data is a critical piece to enable the best results when fine tuning a model. Since Mattermost manages the collaboration platform, the company has a material data advantage to additional fine tuning.

“You can train on what people are saying, but in Mattermost you understand who that person is,” said Tien. “There’s all the digital body language that happens, in addition to the conversation, and we believe that that will provide a material advantage to the model trained with the context of the conversations.”

Technology businesses are built on transitions, said Tien, and Mattermost is a part of the transition to a world that’s using more open-source technology. It’s a transition where the world also cares more about security and privacy, while viewing communications as a mission critical function. Generative AI adds to that transition, providing an opportunity for productivity acceleration.

“What’s really ahead is working with the generative AI community on a framework for data portability so that customers can get the most out of AI while keeping control of their systems and data,” said Tien.