Meta is reportedly testing proactive features for its AI chatbots that let them send messages to users without being prompted, according to Business Insider.
The new feature is part of Meta's internal "Omni" project and is aimed at boosting user engagement, according to training materials from Alignerr, the data provider working with Meta. The chatbots are created using Meta AI Studio, a platform that has allowed users to design custom digital personalities with memory and specific behaviors since 2024. With the new proactive capability, bots can reference past conversations or ask topic-related questions to keep users engaged.
There are strict limits on when a bot can send a message. A bot is only allowed to reach out if the user has sent at least five messages within a 14-day period. Even then, the bot can send just one follow-up message - if the user doesn't respond, the conversation ends there. One example from the training materials: "I hope you had a pleasant day! Have you discovered any new favorite soundtracks?" The bots are designed to pick up on details from previous chats and maintain a consistent personality.
According to Business Insider, freelance workers at Alignerr use an internal Meta tool called SRT to simulate and evaluate the quality of these proactive messages. Reviewers make sure the responses are relevant, friendly, and follow the rules. The training materials stress that bots should avoid controversial or emotional topics unless the user brings them up first. The goal is to position chatbots as reliable, personalized conversation partners.
Meta is looking to turn generative AI into a revenue stream
Meta CEO Mark Zuckerberg has previously described digital agents as a possible way to address loneliness. While the company admits chatbots won't solve the problem, Meta claims they can at least encourage more meaningful interactions. The business case is clear: court documents show Meta expects generative AI products to bring in $2 to $3 billion in revenue by 2025. Letting bots reach out proactively could increase the amount of time people spend chatting, which in turn boosts the platform's value.
Meta declined to tell Business Insider which of these features are already available to the public. Some are apparently still in early testing, while others are being trialed internally.
User retention remains a tricky issue in the chatbot world, with evidence that extended interactions can sometimes have negative effects. Companies like Character.AI have even faced lawsuits over the impact of their chatbots.
