How to Post Real Estate Content to Facebook Daily Using AI Automation

The single most consistent thing you can do for your real estate business on social media is show up. Not with perfect content. Not with viral posts. Just with relevant, professional content that keeps your name in front of people who might need an agent sometime in the next 12 months.

Most agents don’t do this consistently because it takes time they don’t have. Coming up with something to post, writing a caption, finding an image — it adds up. So they post when they remember, which means they don’t post consistently, which means they don’t stay top of mind.

I built an automation that solves this. It runs on autopilot, posts to my Facebook page every weekday, pulls in real estate news from my local market, and writes the caption for me. The total cost to operate it: under $5 per month, and that’s being generous.

Here’s the full setup.

What the System Does

The workflow runs inside Make.com, a free automation platform. At 11:15 a.m. Central every weekday, it does the following:

1. Pulls the latest article from a custom RSS feed of Austin real estate news
2. Sends the article URL to Perplexity AI to summarize it
3. Sends the summary to ChatGPT to write a Facebook post
4. Posts the text and article link to my Facebook Business Page

No manual input. No reviewing before it posts. It just runs.

When I want to check results, I see a post on my page like this one from a recent run: “Thinking about selling your multifamily property? With median prices holding steady, now might be the right time to evaluate your options. [Article link]” — with the article’s preview image populating automatically from the link.

That post reaches my followers, reinforces that I’m an active professional with a pulse on the local market, and takes up more visual space in the feed because the article link generates a preview card with an image and headline.

The Tools and What They Cost

Make.com — free tier. You get a set number of operations per month, and this workflow uses roughly one operation per step per day. For five days a week, you’re well within the free limit.

rss.app — free. You use it to create a custom RSS feed from any news source. I went to the site, searched “Austin real estate,” and it generated a feed URL pulling from relevant local publications. You search for your market’s local news sources and do the same.

Perplexity AI — you need an API key. Sign up at perplexity.ai, go to your account settings, find the API section, and create a key. You’ll need to deposit money to activate API access — I recommend starting with $5. The model I use is Sonar, their cheapest option. It’s more than adequate for summarizing a news article. At the rate this workflow runs, $5 will last you close to a year.

OpenAI (ChatGPT API) — same process. Go to openai.com (not chatgpt.com), log in, go to the API platform, and create an API key. Deposit $5 or $10. I use GPT-4o for the caption writing step — it’s the right balance of quality and cost. The cheaper mini model would also work.

The reason I use both Perplexity and ChatGPT rather than just ChatGPT for everything: Perplexity does a noticeably better job at summarizing articles without making the summary too long or too fluffy. It gives ChatGPT cleaner material to write a caption from.

Setting Up the RSS Feed

The feed is what determines what news your automation reads. At rss.app, you create a new feed by typing in a topic or publication name. I searched “Austin real estate” and got a feed that aggregates relevant articles from local business and real estate publications.

Copy the feed URL that the site generates. That URL goes into Make.com as the trigger — the thing that kicks off the automation whenever a new article is available. If you want to change your news source later, you just swap the URL.

Configuring Perplexity in Make.com

Inside Make.com, the Perplexity module needs your API key and a prompt. The prompt I use: “Summarize the article at this URL in a way that makes it useful for agents to share something valuable with clients.”

The article URL comes from the RSS feed as a variable — when you click on the input field in Make, you can see all the available variables from the previous step and select the URL.

The settings that matter:

• Model: Sonar (cheapest, sufficient for summaries)
• Temperature: 0.7 (adds some natural variation without going off the rails)
• Max tokens: 400 (keeps the summary concise)

Configuring ChatGPT in Make.com

The ChatGPT module takes the Perplexity summary and writes the final Facebook post. The prompt I use instructs it to write as a social media ghostwriter for a real estate agent — professional, direct tone, no emojis, no AI buzzwords. I specifically tell it to avoid phrases like “let’s dive in” and “this revolutionary” because those phrases read as obviously AI-generated.

I also include this instruction: include a link to the article in the post. Here is the URL.

The reason for that last instruction is visual. When Facebook sees a URL in a post, it automatically generates a preview card — a large image pulled from the article, with the headline and site name. That card takes up significantly more space in the feed than text alone. Your post becomes harder to scroll past.

The API model is GPT-4o. Max completion tokens: 5,000 (more than enough for a Facebook post). The variable from the Perplexity step feeds directly into this prompt.

Connecting to Your Facebook Page

The final step is the Facebook Pages module in Make.com. You connect it to your Facebook account, select your Business Page, and map the ChatGPT output to the post message field. The article URL goes into the link field.

You need a Facebook Business Page, not a personal profile — Business Pages support the API connection that Make.com requires. If you’re running a real estate practice and posting about real estate, you should have one anyway.

The posting schedule is set inside Make.com as a cron trigger: weekdays, 11:15 a.m. New York time. Adjust to whatever time works for your audience.

What the Output Looks Like

A recent test post: the automation pulled an article about multifamily investment in Austin, summarized it through Perplexity, had ChatGPT write a two-paragraph post with a call to action, and posted it with the article link. The post looked professional, read like a person wrote it, and the article preview card made it visually prominent in the feed.

The whole process took about two minutes from trigger to live post.

Where This Goes Next

The baseline workflow posts news content every day. That’s useful. But there are several optimizations worth building on top of it.

The one I’m most interested in: routing the same article to a blog post on my real estate website. The workflow already summarizes the article. With one extra step, it can also write a short blog post and push it to my site. Now the same piece of content shows up on Facebook and on my site, and I can track who visits my site from Facebook.

Another optimization: instead of just news summaries, route the automation to post about price reductions and new listings from your MLS. The caption becomes “Here are homes in Austin that reduced their price this week” with a DM call to action for anyone who wants the full list. That’s a lead generation post, not just a brand awareness post.

The full blueprint for the baseline Facebook automation — including the prompts, the Make.com module setup, and instructions for connecting your accounts — is available as a free download (linked in the video above). You download it, import it into Make.com, add your API keys, and it runs.

For the Instagram version of this same workflow, see the Instagram automation post. The setup is nearly identical with a few platform-specific adjustments.

The tools I use across all of these automations are documented on the tools page. If you want to follow along as I build out the next layers, subscribe to the newsletter.