Self-hosted, open source solution enables AI-augmentation of mission-critical operations while maintaining full data control and avoiding vendor lock-in
Palo Alto, CA, June 07, 2023 (GLOBE NEWSWIRE) -- Mattermost, Inc., the leading secure collaboration platform for technical teams, today announced a new "Customer-Controlled AI Architecture" enabling defense, government and technology enterprises to apply generative AI technologies to enhancing or replacing operational workflows while meeting stringent data control requirements and avoiding lock-in to vendor-specific platforms.
"If you're part of an organization that has clear competitors or adversaries, then you're in a race right now to unlock the power of AI for gaining speed and decision advantage," said Ian Tien, Mattermost CEO. "Our release today serves organizations competing in that race who also need to meet strict security and sovereignty requirements."
The open source, self-managed solution provides feature-rich, chat-based collaboration across web, desktop and mobile experiences to engage with colleagues as well as AI bots powered by an interchangeable set of open source AI platforms, known as large language models, or "LLMs". The system can be fully controlled by a customer within their private network or data center, and can even be deployed to air-gapped networks disconnected from the internet.
Benefits of Mattermost's customer-controlled AI platform include:
- Fully-featured chat-based collaboration including 1-1 and group messaging across, web, desktop and mobile, with file and media sharing, search, integrations, custom emojis and emoji reactions, syntax highlighting and custom rendering.
- Conversational AI bots that can be added to channels and engaged like human users to respond to questions and requests based on different LLMs that can be downloaded and run as part of the framework, including models from the HuggingFace AI community.
- Scalable AI model framework that can scale up to deploy on a private cloud or data center using large and powerful open source LLM models for group work, or scale down to run on a commodity laptop, without the need for specialized hardware required by typical AI models, for individual developers to prototype and explore LLM capabilities.
- Conforming security and compliance platform that can accommodate a broad range of custom security and compliance requirements, including automated scanning of the platform's open source code base, monitoring and securing of all incoming and outgoing network traffic and deployment to restricted networks.
This new framework is an initial release in a series of reference architectures that Mattermost, Inc. intends to offer.
"While open source LLMs are in their infancy, we see them on course to materially accelerate operational collaboration," adds Tien. "They'll increase the attention and focus on our most impactful discussions, while reducing the drudgery of sifting through daily information flows. While we are starting with an open source, private cloud framework, over time we intend to add integration with proprietary, public cloud based AI models, because we believe enterprises are best served with a range of solutions to meet different needs."
The Mattermost AI Framework is available for download on GitHub's open source platform at https://github.com/mattermost/mattermost-ai-framework.
About Mattermost
Mattermost provides secure collaboration for technical and operational teams that need to meet nation-state-level security and trust requirements. We serve technology, public sector, and national defense industries with customers ranging from tech giants, to the U.S. Department of Defense, to governmental agencies around the world.
Our self-hosted and cloud offerings provide an extensible hub for technical communication across web, desktop and mobile for incident collaboration, operational workflows, integration with Dev/Sec/Ops and in-house toolchains, and leading unified communications platforms.
We are built on an open source platform vetted and deployed by the world's most secure and mission critical organizations, that is co-built with over 4,000 open source project contributors who've provided over 30,000 code improvements towards our shared product vision, which is translated into 20 languages.
To learn more, visit www.mattermost.com.
Mattermost and the Mattermost logo are registered trademarks of Mattermost, Inc. All other trademarks are the property of their respective owners.
Attachment
- Mattermost announces Customer-Controlled AI Architecture (https://ml.globenewswire.com/Resource/Download/b3142f14-5616-4369-8306-aa5cd1c87bc1)
