All it takes is five minutes on LinkedIn or YouTube to stumble upon the latest tech evangelist prophecy: “How I automated 100% of my business using Make and artificial intelligence!” Listening to these claims, the combination of the automation platform (formerly Integromat) and language models like GPT-4 sounds like pure magic. With a few drag-and-drops and two well-worded prompts, businesses allegedly free themselves from tedious tasks, ready to collect passive income in the sun.
The reality, obviously, looks much more like an engineering nightmare than a productivity miracle.
The No-Code Illusion: A Visual Bloatware Nightmare
Make markets itself as an accessible alternative for non-developers. But beneath its playful canvas filled with glowing icons and blinking modules, the platform hides absurd complexity. What marketing pitches as “No-Code” quickly turns into an unreadable bowl of visual spaghetti.
Building a simple scenario is a total uphill battle:
- Error management: virtually non-existent by default.
- Data complexity: painful manual handling of JSON objects, arrays, and collections.
- Internal logic: expressions with old-school clunkiness.
Where three lines of clean Python code would solve the issue neatly, Make forces users to stack twelve modules, four filters, and three iterators. The result? Instead of writing code, teams spend their days debugging incomprehensible visual nodes.
Injecting AI: Strapping a Rocket Engine to a Cart
Integrating AI into Make is the peak of the charade. Promises of intelligent workflows result in practice in a blind cockpit hooked up to a random number generator.
Language models (LLMs) remain inherently unpredictable: they hallucinate, change response formats without warning, and sometimes miss the most basic instructions. Injecting this unpredictability into the heart of Make scenarios—which are by definition rigid and strict regarding data structure—is a guaranteed recipe for disaster.
- Day 1: the workflow perfectly summarizes an email and sends it to Slack.
- Day 2: the AI adds one extra quote mark or responds in Spanish, breaking the next module and crashing the entire pipeline.
Do not rely on Make to alert users clearly: the crash is only discovered when customers complain or when operations have been stalled for 48 hours.
The Heavy Price Tag of Modular Automation
Financially, Make’s business model relies entirely on the number of “operations.” Every module activation incurs a cost. As soon as AI enters the mix, the volume of operations skyrockets: validating responses, re-parsing, multiple retries on errors, splitting text into chunks. The free or entry-level tier melts like snow in the sun during a single afternoon of testing.
On top of that comes the bill for the OpenAI or Anthropic API running in the background. The supposedly “cost-effective” solution ends up costing more than a dedicated, proven SaaS subscription or hiring qualified technical support.
Endless Maintenance and Surrendered Sovereignty
The ultimate trap of the Make + AI pairing lies in the illusion of “set it and forget it.” A Make scenario never stays stable. A third-party API update, a model change at OpenAI, or a schema modification in Google Sheets is enough to bring everything crashing down. The user does not become a time-liberated entrepreneur, but a full-time mechanic servicing a fragile, leaky pipeline.
Finally, there is the strategic issue: entrusting all corporate data, emails, and customer workflows to a chain of opaque third-party dependencies creates major risks for security and GDPR compliance.
Verdict on Make and AI Automation
Make and AI are not useless tools, but their marketing often relies on heavy overpromising. A cobbled-together assembly of Make modules and AI prompts will never replace clean software architecture or a genuine business strategy. For a few isolated tasks, the solution works fine. To build the operational core of a business, it guarantees building a house of cards on top of a volcano.
Cet article est rédigé par Blaise Grenier, Journaliste reporter d'images et auteur. Découvrir le parcours de l'auteur et la charte éditoriale »