Main information
- The Meta AI studio is developing proactive chatbots to improve user engagement.
- These chatbots will initiate conversations and monitor the interactions spent in a 14 -day window after the initial interaction.
- Interactions with chatbots must comply with Meta’s content and responsibility standards, avoiding any sensitive or harmful content, unless the user takes the initiative.
Meta develops customizable chatbots within its AI studio platform, designed to be more proactive in interaction with users. Internal documents reveal a project whose code name is “omni” and which aims to improve user retention by allowing these chatbots to start conversations and follow past interactions. This reports Business Insider.
The objective is to bring added value and encourage users to return to the platform. Although it is not designed as a solution to loneliness, Meta hopes that this functionality will increase the commitment with its AI Studio, accessible both independently and through Instagram. Launched in the summer of 2024, AI Studio allows everyone to create personalized chatbots with unique personalities and memories.
Chatbot customization
The training guidelines for the Data Labeling Society Alignerr illustrate how a persona called “The Masestro of the Magic of Cinema” could send proactive messages to users: inquire about their new favorite soundtracks or offer recommendations for cinema evenings. Meta confirms having tested the sending of monitoring messages for AI in her studio, stressing that these followers only occur after a user started a conversation and that they are limited to a 14 -day window after the initial interaction.
Users can create highly personalized chatbots, such as chefs suggesting recipes or interior architects offering decoration tips. Creators and influencers can use these bots to manage interactions with fans on Meta platforms.
Responsibility and standards
Meta insists that her AI will not contact users persistently if there is no response to the first monitoring message. In addition, chatbots must join Meta content and responsibility standards, avoiding sensitive or harmful content, unless the user speaks about it explicitly.
Freelancers use an internal revision tool called SRT to simulate conversations with bots, assess proactive messages and refine the text to align it with Meta guidelines.
Conversations adapted to your interests
The emphasis is placed on personalization and taking into account the context, ensuring that each interaction with the chatbot reflects the persona attributed to it and the history of previous conversations. Messages must be positive, relevant and avoid controversial or emotionally heavy subjects, unless the user has taken the initiative. The best messages refer to concrete details from previous interactions, thus creating a more natural and engaging experience.
