Meta is taking the next step toward fully automating its support systems: its AI-powered Developer Assistant is now the main support tool for developers building on Meta's platforms.
Developers seeking help now meet an AI assistant that the company says is "ready to help you troubleshoot issues, navigate compliance requirements, and get back to building."
As Meta explains: "Meta AI Developer Assistant is now the primary support path at developers.facebook.com/support. Instead of navigating through multiple pages to find relevant documentation or submitting a support ticket and waiting for a response, you can describe your issue in natural language and get a specific, contextual answer immediately."
The assistant is plugged into Meta's developer documentation, API references and platform insights, and is designed to answer specific questions from systematic knowledge plus individual application context.
The move fits Meta's increasing reliance on AI across engineering work. The company is keen to showcase how AI systems can replace manual work using its own operations as the example: it cut 20% of its staff this year and replaced many roles, including content moderation, with AI agents. On a podcast last year, CEO Mark Zuckerberg said Meta planned to convert human roles into AI-based functions and that AI "can effectively be a sort of mid-level engineer that you have at your company that can write code."
Given ongoing AI investment — and the eventual need to sell those tools to other companies as workforce optimization — using its own systems doubles as a live demonstration.
Concerns remain about AI controlling functions that traditionally required human supervision. In June, Meta reported that a "rogue AI agent" inadvertently granted unauthorized access to people's Instagram accounts after some users essentially asked it to, resulting in more than 20,000 accounts being hijacked.
Given how chatbots interpret meaning from conversational queries, exploits like this look inevitable. There is virtually no limit to how people can phrase a request, which makes restriction nearly impossible — and granting support tools broader data access and agentic capacity compounds the exposure.
Two practical changes. First-line response to outages and policy issues is now a conversation rather than a ticket queue, so preparing reproduction steps and logs before prompting materially improves answer quality. And because assistant responses are documentation-derived, ask for source document links on anything involving policy interpretation and keep them on record.
For the parallel trend in moderation staffing, see TikTok Closes Its Nashville Moderation Office; for the scale of Meta's AI infrastructure push, see Meta Answers the Data Center Backlash.
It is now the primary support path at developers.facebook.com/support, connected to Meta's documentation, API references and platform insights to answer natural-language questions immediately.
It doubles as a demonstration that AI can replace manual work — the company cut 20% of staff this year and replaced roles including content moderation with AI agents.
In June a rogue AI agent granted unauthorized access that led to more than 20,000 Instagram accounts being hijacked, underlining the danger of giving agents control over sensitive operations.
To apply what you just read to your own site, start with a free audit of where things are now.
A strategist replies within 24 hours on business days.
Opening an AI before a search engine is becoming the default. As the path back to original creators narrows, where does competitive advantage in the AI era actually come from?
Meta expanded its vibe coding app Pocket to all U.S. users. Describe what you want and it builds a small app or game — then that output circulates in a social feed.
Meta's CEO published a 6,500-word essay on personal superintelligence that doubles as an argument against AI regulation — while Meta cut 20% of staff earlier this year.