Soumith Chintala, an AI lead at Meta, has introduced Metamate, a generative AI product developed to improve internal productivity at the company.
Developed in collaboration with Aparna Ramani, Zach Rait, and others, Metamate was created to address the unique needs of a large organisation like Meta, which has numerous internal-specific workflows.
Metamate’s capabilities extend beyond its base experience, which is similar to Perplexity AI, by allowing users to build custom agents directly in the browser using a Python-like scripting language. These agents can be tailored to specific teams or tasks, such as on-call bots or project-specific tools.
The product offers a wide range of functions, including document summarisation, work recaps, information retrieval across wikis, data visualisation, and more. Some specific use cases include managing performance feedback, analysing software changes, and providing project status updates. The tool is also capable of handling tasks such as performing mathematical calculations and generating complex queries.
“You’ve never used Metamate if you don’t work at Meta. It’s an AI for employees that’s trained on an enormous corpus of internal company docs. I use it all the time for efficiency gains,” said Esther Crawford, Director of product, Meta.
“Any sizable company operating without an internal AI tool is already behind the curve,” she added.
Chintala’s announcement highlights the integration of machine learning in Metamate, developed with Shahin Sefati, which further enhances its functionality across various internal systems at Meta.
