Make Review: My Honest opinion After 5+ Years of Automation 2026

I still remember the first time I opened Make (back when it was called Integromat). I was staring at a blank canvas, a little intimidated by all the circles and lines, wondering if I’d made a mistake switching from something simpler.

Fast forward to 2026, and that visual canvas has become my secret weapon. I’ve used it to automate processes for several small businesses—synchronizing CRMs, managing lead flows, building internal dashboards, and even dabbling with the new AI agents.

But let’s be real: it’s not all sunshine and roses. The platform has evolved a ton in the last year alone, but some of the old frustrations remain. So, if you’re wondering whether Make is the right fit for your team in 2026, here is my unfiltered take.

What is Make?

At its core, Make is still the visual automation platform that connects your apps and lets you build complex workflows without code. You drag modules onto a canvas, link them together, and watch your data flow from left to right. It’s like drawing a flowchart that actually does work.

But the Make of 2026 is different from the one I started with five years ago. The biggest change? AI is no longer an add-on; it’s baked into the canvas.

Earlier this year, Make rolled out the next generation of its AI Agents . This isn’t just a ChatGPT module anymore. You can now drop an “Agent” right onto your canvas. This agent can make decisions, use tools, and process multi-modal inputs like PDFs and images, all while showing you its “reasoning” in a side panel . It’s a move to stay ahead of competitors, and honestly, it’s pretty cool to watch an AI logic gate decide where to route a support ticket.

So, Make is no longer just the “power user’s Zapier.” It’s positioning itself as a visual operating system for both deterministic automation and agentic AI. Whether that’s overkill for your small business or exactly what you need depends on your appetite for complexity.

What I Still Love About Make

Despite the new bells and whistles, the core strengths that kept me here for five years remain rock solid.

1. The Visual Canvas is Still Unbeatable

I’m a visual thinker. Looking at a list of “if this then that” statements in other tools makes my brain hurt. Make’s canvas lets me see the entire symphony at once.

When a colleague or client asks, “How does the lead get from the website to Slack?”, I don’t have to write a novel. I just screenshot the canvas. You can see the trigger, the filter that checks for spam, the router that splits B2B and B2C leads, and the final actions. This clarity saves hours of explanation and makes maintenance a breeze. With the new AI agents sitting right there on the canvas, their “thinking” is transparent too—you can see exactly why it took a certain path .

2. The New AI Agents Are a Game Changer (When They Work)

Look, I was skeptical about “AI agents”—it felt like buzzword soup. But I tested the new feature by building a simple “Customer Support Triage” agent.

I connected our help desk email to Make. The agent receives the email, reads the content, and decides: “Is this a billing issue, a technical problem, or a sales question?” Based on its confidence, it either sends a personalized reply, creates a task in Asana, or escalates to a human. It took me about 30 minutes to set up using one of their new Library of Agents templates .

For a small team, this is like hiring a virtual assistant for a few dollars a month. The “in-canvas chat” to refine its behavior is brilliant—you can literally ask it why it made a decision and tweak its instructions on the fly.

3. Unmatched Flexibility for Complex Logic

Make has always been the king of complex scenarios. Need to iterate through a list of 100 JSON items, filter out the ones with null values, aggregate them back into a CSV, and send it to three different places? Make handles this natively with its iterators, aggregators, and routers.

This flexibility means you rarely hit a “wall” where the tool can’t do what you need. If there’s no native module for an app, the HTTP module lets you call any API directly . You have control.

4. Data Transformation is Smoother

Handling data between steps used to be a bit of a puzzle. In 2026, the mapping experience is smoother. When you click on a module field, Make shows you exactly what data is available from previous steps in a clear, searchable menu. It handles arrays and bundles more intuitively, which cuts down on the trial-and-error time significantly.

Where Make Still Frustrates Me

Okay, time for the tough love. For all its power, Make has some sharp edges that have drawn blood (metaphorically… mostly).

1. The Pricing Still Feels Like a Puzzle

Make uses an operation-based model (each module execution = 1 operation). On paper, the plans look great. But in practice, your credit consumption can be a bottomless pit .

A real-world example: I set up a simple scenario to watch a Google Sheet for new rows and send a Slack message. That’s 2 operations per run. 500 runs a month? 1,000 ops. Fine. But if I add a filter to check if the row is complete, that’s a third operation. If I add a module to format the date, that’s a fourth. It adds up scarily fast.

And unlike Zapier, where filters and formatting steps are free, everything in Make costs an operation—including checking for new data (polling), even if nothing is there . This “hidden cost” of logic makes it hard to forecast your bill. You can easily blow through 10,000 operations without doing anything that feels “heavy.” Competitors like Latenode are starting to challenge this model by charging for execution time rather than per step, which is tempting for complex workflows .

2. The Learning Curve is a Cliff, Not a Slope

Make is marketed as “no-code,” but that’s only half true. It’s no-code for simple tasks. The moment you need to do something mildly complex, you have to understand technical concepts.

You need to know the difference between an array and a bundle. You need to understand what JSON is. You need to grasp how APIs handle data. The new AI agents add another layer of complexity—prompt engineering and understanding “tool use.”

For a non-technical business owner, this can be paralyzing. I’ve spent countless hours debugging a scenario, only to realize the problem was a stray comma in a data bundle. If you don’t have someone on your team who enjoys this kind of logic puzzle, Make can become a time-sink rather than a time-saver.

3. Debugging Can Be a Nightmare

Yes, the execution history is detailed. But when an error happens, it often feels like you’re a detective with no flashlight.

You get an error message like Error: StructureMismatchError: Expected array, received object. Great. Which module? Which piece of data? You have to click through the history, expand the inputs and outputs of every module, and manually trace where the data structure broke. It’s tedious.

Zapier’s error messages and logging are generally more user-friendly . In Make, you often feel like the tool is telling you that you’re wrong, but not how to be right.

4. Not All Integrations Are Created Equal

With over 3,000 apps, the catalog is huge . But depth varies wildly. The integration for a major CRM might only handle basic objects (contacts, deals), but fail to support custom objects or complex search criteria. For anything “non-standard,” you’re back to using the HTTP module, which, as mentioned, requires technical chops.

Meanwhile, Zapier boasts over 8,000 integrations, often with deeper coverage for popular apps . If you rely on a long tail of niche tools, Make might not have the native connector you need.

Visit Make.com

Make Pricing 2026: The Good, The Bad, and The Overage

Make’s pricing structure got a bit of a refresh, and it’s important to look past the headline numbers .

Free: 1,000 ops/month. Great for testing, but you’re limited to 2 active scenarios and a 15-minute minimum interval. You’ll outgrow it fast.
Core ($10/month): Starts at 10,000 ops. This is where you get unlimited scenarios and 1-minute intervals. It’s the entry point for real business use.
Pro ($18/month): Also starts at 10,000 ops but adds priority execution and custom variables. “Priority” actually matters when you need things to run on time.
Teams ($34/month): Again, starts at 10k ops but adds team collaboration features.
Enterprise: Custom pricing for large orgs with SSO and audit logs .

The Catch: The starting price is for 10,000 ops. But as your processes grow, you’ll need to buy more. Make has a “slider” that lets you scale up your ops (e.g., to 40,000, 150,000, etc.) with a corresponding price increase .

For a team of 5 with moderate usage, you could easily land in the $40–$50/month range. For a team of 15 with heavy usage and custom code, expect to pay over $200/month . It’s not cheap, but for the power you get, it’s often worth it—provided you have someone to build efficiently.

Make vs. The World (2026 Edition)

The automation landscape is more crowded than ever. Here’s how Make stacks up today:

vs. Zapier: Zapier is the simplicity king. It’s easier to learn, has more apps, and its task-based pricing (where filters and paths are free) is more predictable . Make wins on visual power, complex logic, and now, arguably, with its more integrated AI agent builder. If you’re a power user, you’ll love Make. If you just want things to work without thinking, stick with Zapier.
vs. n8n: n8n is the open-source champion. If you have the technical chops to self-host, n8n gives you incredible control for free . Make’s advantage is its polished interface, managed hosting, and huge library of pre-built connectors. You pay for that convenience.
vs. Latenode: This is the new kid on the block making waves in 2026. It offers a similar visual builder but with a time-based pricing model and built-in access to 400+ AI models for a flat fee . For complex, AI-heavy workflows, Latenode’s pricing is dramatically cheaper. Make’s advantage is its maturity, larger app catalog, and the new visual AI agent builder .

The Verdict: Is Make For You in 2026?

After five years, I’m not planning to leave Make. It’s the tool that bends to my will, not the other way around. But I go into it with my eyes wide open.

Make is a fantastic choice if:

You have someone on your team who enjoys logic and is willing to climb the learning curve.
You need to build complex workflows with branching, loops, and error handling.
You want to experiment with visual AI agents that are transparent and controllable.
You’re migrating from Zapier because you’ve hit its complexity ceiling.

Make is probably NOT for you if:

You just want to connect two apps with a simple “if this then that” rule (use Zapier or IFTTT).
You have zero technical aptitude on your team and no budget for a consultant.
You need to process massive volumes of data and are worried about unpredictable costs (check out n8n or Latenode first).
You live entirely inside the Microsoft ecosystem (Power Automate is a better fit).

Final Thought: Make has successfully evolved from a scrappy underdog to a mature, powerful platform. The addition of transparent AI agents in 2026 feels like a natural evolution, not a gimmick. It’s a tool for builders. If that’s you, you’ll love the power it puts in your hands. Just be prepared to pay for it—both in dollars and in the time it takes to master.

Frequently Asked Questions

Is Make free?
Yes, there is a free plan that gives you 1,000 operations per month. It’s perfect for testing, but you’ll likely need a paid plan for serious business automation .

How much does Make cost in 2026?
Paid plans start at $10/month for the Core plan, but this only includes 10,000 operations. You can scale up your operation limit, which increases the price accordingly .

Is Make better than Zapier?
It depends. Make is better for complex, visual, multi-step workflows. Zapier is better for simplicity, speed of setup, and predictable task-based pricing .

Does Make have AI features?
Yes. In early 2026, Make launched a new generation of visual AI agents that live directly in the scenario builder, allowing for transparent, AI-driven decision-making within your workflows .

Can non-technical people use Make?
For basic tasks, yes. For anything complex, you will need a solid understanding of logic and data structures, or be willing to invest time in learning