Mindvalley Stories
Case Study · AI Mastery

A boutique hotel owner built a social media system that now publishes on its own

An Airtable calendar wired to Make and Meta now posts to Instagram every night at 9pm, with the right photo, then updates itself.

Running a small boutique hotel in Ibiza, this student used to publish every post by hand, one platform at a time. During AI Mastery they built a working automation that runs without them, after a dozen hours of failed tests and one abandoned tool. It is a clear picture of what persistence and a careful method can produce.

Marie
Runs Pure House Ibiza, a boutique hotel with only a few rooms that hosts weddings and events.

Pure House Ibiza automated publishing system

Before

Pure House Ibiza is deliberately small, and its owner wanted that restraint to show up in every caption. But posting meant doing it by hand, one platform at a time. They knew that rushing into automation without clear foundations would only make things louder, not better.

“Automating publishing without laying that groundwork first is just automating noise, faster.”

In their own words

The turning point

In the AI Mastery program, they started with the brand's DNA: its tone, its posture, and what it refuses to say. Then, with Claude, they built an Airtable editorial calendar, going in circles over messy fields and statuses until a structure held up. A first version with Buffer failed because images had to be added by hand, so they removed it and connected Make directly to Meta's own Instagram and Facebook modules.

“At some point I knew Buffer's error message better than my own Instagram captions.”

In their own words

After

Today the system runs every evening at 9pm. Make fetches the day's post from Airtable, publishes it on the right platform with the right photo through ImgBB, and marks it done. The Instagram branch is tested live, and it was the one project of three they saw through to the end.

What it lets people do

The hotel's editorial calendar now drives publishing by itself, so posts go out on time without anyone opening an app. The owner can plan content once, in one place, and keep the brand's voice consistent across platforms.

What they learned

They learned that AI can get precise technical details wrong, sometimes silently. Their safeguard was simple: share a screenshot at every step before saving anything. That habit, they say, prevented more than one quiet disaster.

“My safeguard was simple; I sent a screenshot at every step of the configuration, before saving anything. It prevented more than one silent disaster.”

In their own words

Pure House Ibiza automated publishing system was built during AI Mastery.

Read the full submission
👀 The starting point brand DNA Before even thinking about automation, I started by clarifying Pure House Ibiza's brand DNA; The tone, the posture, what the brand says and what it categorically refuses to say. Pure House is a boutique hotel with deliberately small dimensions, a few rooms only, we host beautiful weddings and events, and that restraint had to show up down to the smallest Instagram and Facebook captions. Automating publishing without laying that groundwork first is just automating noise, faster. 🤝 Building the database with Claude Once that framework was in place, with Claude's help, I built an entire Airtable table to serve as an editorial calendar: every upcoming post, its date, platform, caption, photo, status. I won't lie; Claude and I went in circles for quite a while before landing on a structure that actually held up, between badly named fields, duplicate columns, and statuses that meant nothing to anyone but me. But we got there, and this database became the backbone of the whole system. 💸 The detour through Metricool Next, I needed this calendar to publish itself. I looked into Metricool, an all-in-one tool that could have handled a good chunk of the automation natively. But the price didn't make sense for what I needed it for. I wasn't about to pay for a premium subscription to publish three olive oil photos and two sunsets a week. 🧨 The first attempt — and Buffer's failure So I built a first version using Airtable, Make, and Buffer for scheduled publishing. Except Buffer imposes strict restrictions on Instagram images through its API; There was no way to pass it a photo URL automatically. I had to open Buffer and add every image by hand, which completely defeated the point of the whole project. I won't sugarcoat it: we went in circles on this one too, for a good while… Easily a dozen cumulative hours of back-and-forth, failed tests, "this should work" followed by "nope, actually it doesn't." At some point I knew Buffer's error message better than my own Instagram captions. 🏆 The final architecture I ended up removing Buffer from the equation entirely, connecting Make directly to Meta's native modules: Instagram for Business and Facebook Pages. These modules accept image URLs directly, via ImgBB. The final scenario runs every day at 9pm: Make fetches the day's post from Airtable, publishes it on the right platform, with the right photo, then automatically updates the status. The Instagram branch has been tested live, successfully; the Facebook branch follows the same logic. 🧐 What I learned One central lesson from this whole journey: AI can get things wrong, even on precise technical details. A misidentified field, a checkbox that silently overwrites existing data. My safeguard was simple; I sent a screenshot at every step of the configuration, before saving anything. It prevented more than one silent disaster. 🎓 Why this project, and not another I'm currently building two other applications as part of this training, both still works in progress. I chose this automation as my graduation project precisely because it's the only one of the three I saw through to the end; Actually tested, working, and running today without me. 🙏🏻 In conclusion I'm very proud of how far this has come. Some weeks ago, I was publishing my content by hand, one platform at a time. Today, my editorial calendar drives publishing on its own. Thank you so much to the whole AI Mastery team for this fantastic journey, I'm genuinely happy I saw it through to the end.

Written 7 August 2026 Submitted by the storyteller

Every story is written by a Mindvalley student or by someone our teachers have worked with. How we collect and check stories

More from AI Mastery

Pure House Ibiza automated publishing system: a AI Mastery case study | Mindvalley Stories