Mattermost now provides a framework for businesses and operations to take advantage of the AI revolution while maintaining control over their data.
OpenOps can help those who want to use AI to analyze and draw conclusions from their own data – without becoming dependent on different technology providers.
The OpenOps framework means that developers and businesses themselves can implement and host a “sandbox”, that is, an environment where you can evaluate opportunities and risks with different AI models with full control. The system can be run completely offline while using self-hosted large language models (LLM) AIs found on Hugging Face, such as Falcon LLM and GPT4All. The system can also be connected to vendor-hosted models such as Azure AI, OpenAI ChatGPT and Anthropic Claude.
Evaluating AI’s benefit to businesses
OpenOps is an open source framework and makes it possible to evaluate different possible uses for AI that are interesting to companies and authorities. These AI models can be used to provide increased efficiency by speeding up workflows and leading to faster processes and decisions.
“Every business is now accelerating its AI use and racing to be the first to increase its efficiency and gain an advantage over the competition,” said Ian Tien, CEO of Mattermost. “We created OpenOps to help businesses evaluate and quickly find potential uses for AI that can improve their internal workflows.”
More and more people see that there are a number of different AI models that are interesting for their business. Unfortunately, it is difficult to use many of these AIs without sharing their data with the providers.
At the same time, there is increased regulation in the area and a risk of data leaks, which means that businesses want to maintain control over their own data and intangible assets.
“We want the AI tools to be controlled by the businesses themselves. We can only do that in a platform based on open source code. Therefore, we now want to show how this becomes possible in practice, by releasing a framework for how businesses can use and experiment with the AI model they want in the way they want. In this way, we help them get the most out of the AI revolution,” says Ian Tien, CEO of Mattermost.
Four steps to OpenOps
The framework for OpenOps recommends that businesses follow four different steps to develop AI use:
- Set up a self-hosted “sandbox”
- Ensure control over data with different AI models
- Investigate security and regulatory compliance for The AI models
- Let users test the solutions