Skip to content
Browse agents

Make Review

by Make (Celonis) ยท Updated

A visual automation canvas with AI agents, priced per operation rather than per task.

Semi-autonomousWorkflow AutomationCloud
4.1
โ˜…โ˜…โ˜…โ˜…โ˜…Editor score
Capability
โ˜…โ˜…โ˜…โ˜…โ˜…4.0
Ease of setup
โ˜…โ˜…โ˜…โ˜…โ˜…4.3
Integrations
โ˜…โ˜…โ˜…โ˜…โ˜…4.6
Value for money
โ˜…โ˜…โ˜…โ˜…โ˜…4.4
Reliability
โ˜…โ˜…โ˜…โ˜…โ˜…4.1
Visit Make

Opens the vendor's site in a new tab

Overview

What does Make do?

Make is a visual automation platform that sits between Zapier's simplicity and n8n's flexibility. Scenarios are built on a canvas where each module is a step, and the visual model handles branching, iteration and error handling far more legibly than a linear step list.

AI Agents were added on top of that foundation: you define an agent, give it a system prompt and attach scenarios as tools it can invoke. The agent decides which to call and in what order, while the scenarios themselves stay deterministic. Same hybrid pattern as n8n, without needing to run a server.

The commercial difference is the billing unit. Make charges per operation โ€” roughly one module execution โ€” rather than per task or per Zap run. For workflows that touch many steps, that usually works out cheaper than task-based competitors, though an agent that calls a lot of tools consumes operations quickly.

Features

Make features

Visual scenario canvas

Branching, iterators, aggregators and error handlers laid out spatially, which beats a linear list once logic gets real.

AI Agents with scenario tools

Attach existing scenarios as tools an agent can call, keeping the deterministic parts deterministic.

2,000+ app integrations

Broad coverage plus a generic HTTP module and a custom app builder for anything missing.

Per-operation pricing

Charged by module execution rather than per run, which usually favours multi-step workflows.

Error handling routes

Explicit error paths per module โ€” the feature that separates a demo from something you can trust overnight.

Make Grid

A visual view of how scenarios interconnect across the organisation, useful once you have dozens.

Integrations

Integrations and how to connect Make

What Make connects to, and the order to set it up in.

  1. Create a scenario

    Start from a trigger module โ€” a webhook, a schedule or an app event โ€” then add modules along the canvas.

  2. Connect apps once

    Each connection is authorised once and reused across scenarios. OAuth for the major platforms, API keys elsewhere.

  3. Build the deterministic path first

    Get the workflow working without AI. Add the agent only where a genuine decision is needed โ€” cheaper and far easier to debug.

  4. Add an AI Agent module

    Write the system prompt and attach only the scenarios it genuinely needs as tools. Over-provisioning tools causes erratic behaviour.

  5. Add error handlers, then schedule

    Set an error route on every module that touches an external service before you let it run unattended.

Fit

Who is Make for?

Great fit

Ops teams wanting more power than Zapier

The sweet spot. Branching and iteration without needing to run infrastructure.

Great fit

Teams with multi-step, high-volume workflows

Per-operation billing usually beats per-task pricing once workflows have real depth.

Good fit

Visual thinkers

The canvas makes complex logic legible in a way linear builders never manage.

Poor fit

Teams with data residency requirements

Poor fit. Cloud-only, with no self-hosting option โ€” use n8n instead.

Use cases

Best use cases for Make

Multi-branch business processes

Approval routing and conditional handling where the logic genuinely forks.

E-commerce operations

Order, inventory and fulfilment flows across Shopify, Stripe and a warehouse system.

Bulk data processing

Iterators and aggregators handle batch work that step-based tools handle badly.

Agent-assisted routing

Deterministic pipeline with one agent deciding classification or destination.

Pricing

Make pricing

Subscription tiers priced in credits (formerly 'operations'). Free tier: Yes โ€” 1,000 credits per month.

Free

$0

Perpetual, not a trial.

  • 1,000 credits/month
  • Full app catalogue
  • No time limit

Core

$9/month

Entry paid tier.

  • 10,000 credits/month
  • Unlimited active scenarios
  • Scheduled control

Pro

Most common

$16/month

Make's recommended tier.

  • 10,000 credits/month
  • Priority execution
  • Custom variables

Teams

$29/month

Collaboration features.

  • 10,000 credits/month
  • Shared templates
  • Role-based access

Enterprise

Custom

Governance and volume.

  • SSO
  • Advanced security
  • Custom volume

Pricing last verified . Vendors change pricing frequently โ€” always confirm on Make (Celonis)โ€™s own pricing page before buying.

User sentiment

What other users say about Make

A summary of publicly available feedback about Make โ€” not paid testimonials.

Users who moved from Zapier describe the canvas as the moment automation stopped feeling constrained, and per-operation pricing as a genuine saving on multi-step work. The two recurring frustrations are that operation counting is unintuitive until you have been billed once, and that the visual model gets cluttered on large scenarios.

What users praise

  • Visual canvas handles branching and iteration far better than linear builders
  • Per-operation pricing rewards deep workflows
  • Error handling routes are properly first-class
  • Very broad integration catalogue

Common complaints

  • Operation counting is confusing until the first invoice
  • Large scenarios become visually unwieldy
  • Agent features are newer and less mature than the core product
  • No self-hosting, so data residency is a hard blocker

We publish individual user reviews only once they have been collected and attributed to a real source. None have been collected for Make yet. Used it? Send us your review.

Sentiment summarised from: Make โ€” documentation, Make โ€” pricing.

Verdict

Is Make worth it?

Pros

  • Best visual model for complex branching logic
  • Per-operation pricing suits multi-step workflows
  • Strong error handling
  • Free tier sufficient for genuine evaluation

Cons

  • Operation accounting is unintuitive
  • Big scenarios get visually messy
  • Agent layer less mature than the automation core
  • Cloud-only
4.1
โ˜…โ˜…โ˜…โ˜…โ˜…Editor score

Assessed from vendor documentation, pricing pages and public user sentiment โ€” not a hands-on trial. How we assess agents.

Try Make

FAQ

Make frequently asked questions

What counts as an operation?

Roughly one module execution. A ten-module scenario running once consumes about ten operations โ€” so a workflow processing 100 records can consume far more than the run count suggests. Model this before choosing a tier.

Make or Zapier?

Make for complex branching, iteration and multi-step workflows where per-operation pricing works out cheaper. Zapier for the widest integration catalogue and the simplest possible setup.

Make or n8n?

Make if you would rather not run infrastructure. n8n if you need self-hosting, data residency, or unlimited volume at a fixed server cost.

Can I build a custom integration?

Yes. The HTTP module covers most APIs, and a custom app builder lets you package a private integration for reuse across scenarios.

Compare

Make vs the alternatives