Meta Accelerates Development of AI Agents: Instagram to Potentially Integrate New AI Shopping Tools

王昱

Original report by Ebrun. May 9th, according to foreign media reports, Meta is accelerating the research, development, and deployment of AI Agents and plans to launch two new AI tools. These tools will enable users to have AI systems perform a variety of tasks on their behalf with minimal human intervention, including shopping, information organization, and marketing operations.

It is reported that one of the AI agent products under development is named "Hatch." This tool will be directly integrated into Meta's application ecosystem and will be capable of autonomously executing a series of complex tasks based on users' conversational instructions.

Based on currently disclosed information, Hatch is not merely a traditional chatbot but is closer to OpenAI's Assistant: it can not only answer questions but also invoke workflows, continuously execute tasks, and automatically adjust its behavioral logic based on real-time data.

For Meta, its true advantage lies in its vast application ecosystem.

Currently, Meta owns several global-scale social platforms including Facebook, Instagram, Messenger, WhatsApp, and Threads, covering scenarios such as social networking, content, communication, and commerce. This also means that once AI Agents are truly deployed, the data and application scenarios they can access will far exceed those of traditional single-application AI assistants.

According to the currently exposed concepts, future users might be able to ask the system to automatically summarize chat content from major messaging apps daily and generate a "Daily Information Recap" email; they could also have the AI automatically track competitors' latest promotional activities to generate market trend summaries; or even have it automatically scan consumer questions about a certain type of product and generate marketing campaign ideas accordingly.

Furthermore, AI may take on more social media management tasks.

For example, automatically sending birthday wishes to contacts, capturing trending topics in real-time, and generating relevant content. As model capabilities and platform data are further integrated, future related application scenarios are expected to expand further.

Compared to information organization and content generation, the market is paying more attention to the AI shopping capabilities Meta is advancing.

It is revealed that Meta is also developing a new AI shopping tool for Instagram. This system will not only help users find products and discount information but will also be able to complete the entire purchase process on behalf of the user.

This suggests that Instagram may evolve further from a "content seeding platform" towards an "AI automated shopping platform."

For instance, when a user sees a T-shirt of interest while browsing Reels short videos, they could directly instruct the AI agent to search for the same product, compare prices across different platforms, filter for the best discount, and ultimately automatically complete the ordering and delivery process, all without the user leaving the current content page.

In fact, Instagram has attempted to integrate AI and e-commerce multiple times before. Whether it was early algorithm-based product recommendations, the Shop Tab shopping entry point, or later features like Meta AI providing product Q&A and recommendations within direct messages, the core logic has always been to improve the efficiency of product discovery for users.

However, unlike previous efforts, the newly exposed tool clearly shows for the first time Meta's intention to have AI directly replace users in completing shopping tasks—the role of AI seems to be gradually shifting from a "shopping advisor" to a "transaction agent."


Ebrun will continue to track and report on this development. To learn more about related information, please scan the QR code to follow the author's WeChat.

[Copyright Notice] Ebrun advocates respecting and protecting intellectual property rights. Without permission, no one is allowed to copy, reproduce, or use the content of this website in any other way. If any copyright issues are found in the articles on this website, please provide copyright questions, identification, proof of copyright, contact information, etc. and send an email to run@ebrun.com. We will communicate and handle it in a timely manner.

Like

Translated by AI. Feedback: run@ebrun.com