Meta is reportedly working on two new agentic artificial intelligence tools, which would enable users to let its AI systems conduct tasks on their behalf, including shopping, with minimal human guidance.
According to The Information, Meta is currently working on Hatch, an AI agent that would be made available within its apps, and would be able to conduct tasks based on conversational guidance.
Hatch sounds similar to Open Claw, another agentic AI tool that goes beyond basic question and answer responses, and can be linked into a user’s workflow to undertake tasks on its own. For example, a business owner could ask Open Claw to ensure that its prices always remain 5% lower than competitors, and the system would check for price variations each day, crossmatch that with listed inventory, and align with its listings accordingly.
These tools represent essentially the next level of AI interaction, where people don’t just ask an AI chatbot for an answer to a question, but they empower it to take action. Because the chatbot is connected to various tools and systems, it can then interpret the query, and proactively undertake those commands with no further input.
In a Meta context, that could enable users to:
- Summarize chats across all of its messaging apps in a catch-up e-mail each morning
- Provide an overview of all new promotions being run by competitors
- Scan for questions being asked about related products and present campaign ideas based on these
- Send happy birthday messages to all users who are celebrating a birthday
- Find trending topics and generate themed posts
These are some quick examples, and there are possibly many more in-depth, smarter applications. But essentially, any activity that users conduct across Facebook, Instagram, Messenger, WhatsApp and Threads could be automated and streamlined to save time and effort in management.
Shopping could also be automated and streamlined.
According to The Information’s intel, Meta is also developing a shopping tool for Instagram, which would enable users to find out more about relevant products and offers, and conduct purchases on their behalf.
This would seemingly enable users to ask the AI agent to monitor for the best deals, and help users find the most relevant product matches, which would then also empower the system to handle the purchase process for each on their behalf.
So if a user is scrolling through Reels and sees a cool t-shirt, they could ask the system, then and there, to find it online, find the best price, then have it sent to them, all without leaving the Reels stream.
This is the next level of AI interaction, and given Meta’s massive investment into AI superintelligence, it makes sense that this would be the next stage of integrating more advanced AI tools.
Meta CEO Mark Zuckerberg has himself been something of a guinea pig for these tools, with Zuckerberg training an AI system on his own day-to-day activities, as a means to optimize and automate his own process.
The challenge in implementation could be that general social media users are not as informed about what agentic tools can do, and how they might be able to train the tools to act on more advanced tasks. Open Claw, for example, takes some trial and error with advanced commands.
But maybe, if Meta can simplify this process, it could be close to making a valuable step toward the next stage of AI interaction.
