Meta is advancing its artificial intelligence (AI) strategy for businesses by introducing an in-house productivity tool called Metamate, according to a Financial Times report.
The tool is designed based on the large Meta language model, Llama, for tasks such as programming, searching, and drafting communications for internal and external use.
Prashant Ratanchandani, vice president of engineering at Meta, said there are no plans to release Metamate for external use. He described the initiative as an effort to develop “the world’s best enterprise assistant” and gain insight into enterprise AI applications.
The move aligns with broader competition between technology companies such as OpenAI and Google, which are working on AI-powered assistants to serve consumers and businesses. The market for artificial intelligence agents, capable of performing complex tasks, is expected to grow from $5.1 billion to $47 billion by 2030, according to MarketsandMarkets.
Meta explores AI monetization opportunities through enterprise products. In October, the company appointed Clara Shih, former chief artificial intelligence officer at Salesforce, to lead its Business AI group. Shih aims to make AI accessible to businesses and monetize Meta’s Llama model through applications that take advantage of its open source nature.
Meta CEO Mark Zuckerberg has integrated AI features across Instagram, WhatsApp, and Facebook, targeting users, creators, and businesses. However, the Financial Times report notes that Metamate currently lacks features found in similar tools from Microsoft and Google, such as participating in meetings, scheduling, and performing tasks independently.
Meta employees are said to be using Metamate to encrypt, access internal data, and prepare for client interactions. The company also uses other models internally, including OpenAI’s ChatGPT. Ratanchandani noted plans to make Metamate more like an AI assistant, but noted that it was too early to assess its impact on productivity.
A recent Goldman Sachs report suggests that widespread adoption of AI tools and productivity gains may not be realized until at least 2027. Oren Etzioni, a professor of computer science at the University of Washington, commented that while internal testing is common, Meta may face challenges in transitioning to an AI system. Competitive enterprise AI vendor.