Custom Facebook algorithm: Use AI controls to tune your feed

Facebook announced a new set of controls called Your Algorithm that will let people add and manage topics they want to see in their feed and Facebook Reels, according to The Verge. Meta says you can mark topics or posts as “interested” or “not interested” and that your recommendations "respond right away" after you make those updates.

What Meta is adding to Facebook

According to The Verge, Your Algorithm gives users direct controls to say what they want to see more or less of, and those controls work both at the topic level and on individual posts. You can access Your Algorithm itself, or use "interested" and "not interested" from the three dots menu on a post. The buttons may also appear below posts. Meta says it may follow up with questions to get more detail about what you liked or disliked.

The Verge also notes that Your Algorithm is not entirely new. Meta introduced a similar feature on Instagram earlier this year, and a feature called "Dear Algo" appeared on Threads, the report says.

How the custom Facebook algorithm is supposed to work

Meta told The Verge it is using more large language models in ranking systems, and that those models help interpret natural language, interests, context, and intent. The Verge quotes Meta saying, "LLMs can incorporate the interests, context, and intent behind natural language." That is how Meta says the system will take plain-language input and adapt recommendations within the same session.

The reporting indicates the focus is on making the recommendation system respond to real time signals from users, whether those signals come from topic selection, single-post feedback, or follow up answers to clarifying questions.

Why this matters to small organizations and teams

If you run a small business, nonprofit, or a public sector account, the change affects how your posts reach people who use these controls. When more users actively tell Facebook what they want to see, ranking can shift faster than before. That matters for reach, because signals users give about interest or disinterest feed back into what the service surfaces in feeds and Reels, according to The Verge.

You do not need to be a specialist to act on this. Simple behaviors by followers, like marking a post as "not interested," could lower the visibility of similar content for those people. Conversely, people marking related topics as "interested" could raise visibility for content that matches those topics.

Practical steps you can take now

• Try the controls yourself, so you understand the messages followers might see when they interact with them. Use the three dots menu on a post and the topic controls to see how the feature behaves.
• Encourage followers to give clear feedback if you want to refine who sees what. Ask a small sample of engaged followers how the feed changes after they try the controls.
• Review your content mix and headlines: if your audience uses these tools, content that closely matches chosen topics will perform differently than before.
• Keep basic account security in place, including strong passwords and multi factor authentication, so controls and content stay under your team’s control.
• If you want a plain-English plan for how AI could affect your organization online, consider an AI roadmap that maps features to business goals and risks, like the one we offer at our AI page.

Things to watch as this rolls out

The Verge report makes clear Meta is positioning this as part of a broader push to add AI into ranking and recommendations. That means two practical things to monitor: first, how often the follow up questions appear and how precise they are, because that affects how well the system interprets user intent; second, whether the real time updates change patterns of reach for particular post types or topics.

Because The Verge reports the feature appears in different places, including under posts and in a central Your Algorithm interface, your audience may encounter the controls in multiple contexts. Track engagement and reach for a few weeks to see if there are measurable shifts.

Final thoughts

The Verge describes Your Algorithm as a way for people to speak directly to Facebook about what they want to see, with immediate changes to recommendations and a heavier role for large language models in ranking. For small teams that rely on Facebook for outreach, the sensible response is to test the feature, encourage clear feedback from your audience, and keep your account and content practices solid. This will let you adapt quickly if the new controls change which posts get seen more often.