Make is a team-oriented AI Office and Efficiency visual AI automation platform, with 3,000+ pre-built application connectors AI Agent, MCP and visual scene orchestration as the core, covering the six major automation fields of IT, marketing, sales, operations, finance and HR.

Make

Make’s core parameters and statistics

Make is a visual AI automation platform under Celonis. It uses "Scenario" as the core orchestration unit and combines triggers, actions, routing and AI nodes into a multi-step workflow through drag and drop. The platform provides unified automated middle-end capabilities for the six business domains of IT, marketing, sales, operations, finance and HR.

Projects	Public Information
Official positioning	The visual AI automation platform
Product form	Visual workflow orchestration + AI Agent + MCP Server
App Connectors	3,000+ pre-built apps
AI integration	400+ AI application connectors (including OpenAI, DeepSeek, Perplexity, etc.)
Customer scale	400,000+ organizations, covering 200+ countries
Team size	350+ people
Compliance Certification	SOC 2 Type II, SOC 3, GDPR, SSO
Community Rating	G2 4.7, Capterra 4.8, GetApp 4.8, Gartner 4.6
Latest capability line	AI Agents, Agentic Automation, MCP, Make Grid
Support Platform	Web, API, Webhook

Billing unit: Make uses "Credits" as the unit of measurement for execution - an action of each module in the scene (such as adding a Google Sheet row, calling the OpenAI API) consumes 1 credit. This billing model is more fine-grained than Zapier's task system and is suitable for high-frequency and low-consumption scenarios. However, under high-volume operations, a credit budget needs to be planned in advance.

Ecosystem Scale: The combination of 3,000+ connectors and 400+ AI integrations means that most mainstream SaaS can be directly connected. For medium and large organizations that need to connect multiple business systems, this reduces the repeated investment in building self-built connectors.

Deployment boundary: Make currently only provides cloud SaaS form and does not support self-hosted or private deployment. Organizations where data sovereignty is sensitive need to factor this into their compliance assessments.

Make’s users and market recognition

Make’s market position is built on two lines: post-acquisition brand restructuring and the infusion of AI capabilities. Official public data and third-party ratings provide verifiable reference dimensions.

Customer scale and coverage: The official page shows that more than 400,000 organizations have adopted it, covering 200+ countries, and customers include BNY Mellon, Bolt, Stellantis and other large enterprises. From the perspective of SaaS automation platforms, this level is second only to Zapier, but it has formed a differentiated positioning in the flexibility of visual orchestration.

Third-party ratings: G2 4.7, Capterra 4.8, GetApp 4.8, Gartner 4.6. The four rating curves are highly consistent, indicating that the user experience evaluation on mainstream evaluation platforms is stable. What needs to be noted is the sampling bias on the rating platform - highly active users are more inclined to rate, while the opinions of users who give up due to complexity are more difficult to be collected by the system.

Industry Recognition: After being acquired by Celonis (the leader in process mining), Make gained the synergy between enterprise-level process automation and process mining. Celonis itself is also using Make AI Agents to reduce annual fee audit costs (official success cases are publicly mentioned). This kind of internal adoption of "bitches eating their own dog food" has reference value for corporate customers.

Benchmark Cases: The official success case page discloses FranklinCovey (saving hundreds of thousands of dollars and releasing hundreds of man-hours), Perk (improving internal operational efficiency) and other cases, covering different scales and industries. The prerequisite for implementation is that the organization has a repeatable process that can be standardized.

Make’s cost advantage

• C-side/Individual: Usually a free version is provided to experience the core functions, and high-frequency use requires a paid package subscription.
• API/Developer: Billed by call volume, suitable for development teams that can be flexibly integrated into their own systems.
• Enterprise/Privatized: Contact the business owner for customized quotation and deployment plan. The specific price is subject to the official real-time pricing page.

Main functions of Make

Make's ability is not to simply redo the Zapier model, but to superimpose an AI-native execution and governance layer on top of the bottom layer of visual orchestration.

• Visual scene orchestration: Combine triggers, actions, routers, filters and aggregators into multi-step workflows by dragging and dropping. The router supports branch conditions, sub-scenario calling and loop iteration, and is closer to traditional programming process control in complex logic than Zapier's multi-step Zap.
• AI Agent and Agentic Automation: The AI ​​Agent capability launched in 2025 supports the construction of autonomous agents through natural language prompts or drag-and-drop methods, allowing agents to complete task orchestration in 3,000+ applications instead of single-step execution. The difference from fixed scenarios is that Agent has decision-making capabilities and can dynamically adjust subsequent actions based on intermediate results.
• MCP Server Connection: Make supports incorporating MCP (Model Context Protocol) Server into the automation link, allowing the AI ​​model to directly call external system capabilities through standard protocols. This can reduce duplication of integration for teams that already have MCP governance structures in the enterprise.
• Make Grid panoramic view: A global automated dashboard for enterprise management, showing cross-department and cross-scenario operating status, failure rate and credit point consumption, solving the automation "black box" problem - this is a governance capability that competing products such as Zapier have not yet equally emphasized.
• 400+ AI application integration: Covers mainstream model vendors such as OpenAI, DeepSeek, Perplexity, Claude, etc., making LLM calls a standard node in the scene rather than an external script. The platform does not limit which model to use, users only need to bring their own API Key.

Functional synergy: AI Agent + MCP + Visual Orchestration When the three are combined, the capabilities are not simply added together. A typical example is: the user uses prompt words to describe a business process (Agent construction), the Agent automatically calls the MCP connector to obtain external data, and then performs subsequent approval, notification and write-back through visual scenarios - the entire process evolves from "hard-coded process" to "goal-driven execution".

Make version evolution

Make's version history clearly reflects the three evolutionary stages from independent tool → platform → AI native.

Phase 1: Integromat (~2015 — 2022)

It enters the market with visual multi-step scenario orchestration. The core difference lies in supporting advanced process control such as branch (Router), aggregation (Iterator) and sub-scenario (Sub-scenario). The early user groups are concentrated in operations and data teams that require complex logic. Compared to Zapier’s linear multi-step Zap, Integromat always leads in orchestration flexibility, but at the cost of a steeper learning curve.

Phase 2: Renamed to Make (~2022)

After being acquired by Celonis, the rebranding was completed, and the interface shifted from Integromat's engineering style to a more modern product design. At the same time, Make Grid panoramic view, enterprise-level SSO and audit logs were introduced, and the customer base expanded from individuals/small teams to medium and large enterprises. The core change in this stage is the upgrade of product positioning from "automation tool" to "automation management platform".

The third phase: AI native (~2024 to present)

The embedding of AI nodes (2024) makes LLM inference a standard module in scenarios; the launch of AI Agent (2025) further pushes the orchestration paradigm from "predefined process" to "goal-driven agent". The addition of MCP support provides a standardized protocol channel for connecting AI models to enterprise systems. It is still in a high-frequency iteration period. It is recommended to use a mature version stably instead of following every update.

Make’s technical advantages

Make’s technical advantage lies not in the uniqueness of the underlying infrastructure, but in the depth of matching between the execution engine’s native support for complex workflows and visual governance.

Execution engine vs linear process: Different from Zapier's linear multi-step execution model, Make's scene engine natively supports branching (Router), aggregation (Iterator), sub-scenario calling and error handling branches. This architecture allows it to express common logic in traditional programming such as "if A then B otherwise C, and each of D executes E" without resorting to external scripts. Causal chain: Stronger process control capabilities → Complex business logic can be completed within one scene → Reduced maintenance costs for multi-scenario splicing.

Engineering design of visual orchestration: Make's drag-and-drop interface is not a simple graphical packaging of code, but a unified model of triggers, actions, routing, aggregation and AI nodes into executable DAG (directed without graph). Each node can independently set input and output mapping, error retry and conditional branching. This design allows non-R&D roles to maintain medium-complexity automated processes.

Differentiation at the governance layer: Make Grid integrates scenario execution status, credit point consumption, and failure distribution into a management-readable dashboard, and cooperates with audit logs and permission roles to upgrade automation from "personal efficiency tools" to "department-level infrastructure." For IT governance teams, this means no manual inspection of the running status of each scenario.

Technical limitations: Not supporting self-hosted/private deployment is a common limitation of all SaaS automation platforms, but Make may be excluded in enterprise compliance scenarios. In addition, the relationship between execution delay and credit point consumption in high-frequency execution scenarios is not made public and transparent, and actual performance needs to be verified through PoC before purchasing.

How to use Make

Make provides two interactive entrances, web-side construction and API/Webhook, covering the complete link from drag-and-drop construction to programmatic triggering.

How to use	Applicable stages	Core operations	Prerequisites
Web client construction scenario	Design, testing and operation and maintenance	Drag and drop to configure trigger → action → AI node → routing	Register an account and start the Free plan
API trigger	Production integration	Start and stop scenarios, query status through REST API	Core and above solutions
Webhook trigger	External system linkage	Receive external system request as scene trigger	All solutions
Make AI Agent (prompt word)	Rapid prototyping	Use natural language to describe the target, and the Agent will automatically disassemble and arrange	The current official availability shall prevail

Typical Getting Started Path: After registration, select a scene close to your needs from the template library (officially provides a large number of pre-built templates) and start cloning and modifying it, which is more efficient than building from a blank scene. It is recommended to use the Free solution to run through 1-2 low-risk processes (such as form data synchronization, email notification) in the first week, focusing on verifying the availability of the connector, whether the trigger interval meets the requirements, and whether the failure retry strategy is acceptable.

Production deployment tip: The error handling branch (Handle Error) and notification alarm must be set up before the scene goes online to avoid silent failure. For scenarios involving data writing, it is recommended to first use Dry Run or test context to verify the output format, especially when using AI nodes - uncontrollable LLM output may cause abnormal downstream field formats.

Make’s product pricing

Make uses "credit points" as a unified unit of measurement, and its pricing is open and transparent. However, high-usage scenarios require marginal cost deduction.

Plans	Monthly Fees	Credits/Month	Minimum Execution Interval	Key Differences
Free	$0	1,000	15 minutes	Basic experience, suitable for prototype verification
Core	$9	10,000	1 minute	Unlocked API, unlimited scenes
Pro	$16	10,000	1 minute	Priority execution, custom variables
Teams	$29	10,000	1 minute	Team roles, template sharing
Enterprise	Custom	Custom	Custom	24/7 Support, Advanced Security, Custom Functions

Credit Mechanism: Each module consumes 1 credit per execution. A 5-step scenario executed 200 times per day = 1,000 credits/day (5×200). This means that the 1,000 points/month of the Free plan are quickly exhausted in real business, and only the Core and above plans have daily usability.

Hidden cost items: (1) Purchase of excess credit points - the unit price after exceeding the monthly quota is subject to the official page; (2) AI node inference fee - calling OpenAI and other models requires the user's own API Key, Make does not pay for it; (3) Enterprise solutions require business confirmation - SSO, audit logs, private deployment and other terms are not fully displayed on the public pricing page.

Annual payment discount: The official provides an annual payment option (Pay annually), which usually saves about 15-20% compared to monthly payment. The details are subject to the real-time page.

Make application scenarios

Make’s implementation scenarios focus on cross-system, standardizable digital processes, which can be broken down into the following three categories according to job roles:

• Marketing Automation: cross-platform content distribution (social media → CMS → email), lead scoring and circulation (form → CRM → sales notification), and advertising platform data aggregation. The benefits are reflected in reduced manual transcription and lead response times compressed from hours to minutes. Key points to verify: Whether the data refresh frequency and attribution window of the advertising platform connector meet the requirements.
• Sales and Customer Operations: full-link automation from lead warehousing → distribution → follow-up reminders → contract expiration warnings, combined with AI nodes to make customer message summaries and next step suggestions. Verification focus: Whether the inference delay of the AI ​​node will slow down the pace of sales follow-up; when customer data is involved, it is necessary to confirm that data residency is compliant.
• IT and financial process: work order system linkage (monitoring alarms → creating work orders → notifying the responsible person), cross-SaaS account life cycle management (onboarding → assigning permissions → resignation recovery), automation of expense approval. Cost reduction and efficiency increase: Taking IT work order processing as an example, it takes an average of 3-5 minutes to manually create a work order. After automation, it is triggered in seconds. An IT operation and maintenance specialist can save about 2-3 hours by processing 50 work orders every day.
• HR process: Automation of the onboarding process (offer confirmation → system account creation → equipment application → training email), linkage of attendance and leave approval. In this type of scenario, manual approval points need to be set up for sections involving employee privacy data.

Make is suitable for people

Make's multi-modal strategy can serve four types of roles, but the depth of application for each type is different.

• IT and Platform Team: The best match. IT teams need automated, unified governance across departments, and Make Grid's panoramic view and audit logs provide management visibility. Suitable for organizations that already have multiple SaaS systems and require standardized connection management.
• Operations and Marketing Operators: Codeless orchestration lowers the threshold for process construction, but requires users to have process disassembly capabilities - the ability to transform "the business actions I want to automate" into structured thinking of triggers → actions → conditional judgments. This is not something that comes naturally to all operators.
• Automation Artifact and Efficiency Manager: Teams that focus on ROI can use the Free solution to do PoC and use credit consumption data to reversely calculate the cost of a single process. Make’s billing transparency is an advantage in this scenario.
• Enterprise and Compliance Buyers: SOC 2 / GDPR certification + SSO + audit logs, complete compliance basic capabilities. But the not applicable scenarios are also clear: organizations that need a completely private deployment (Make has no self-hosted option); real-time scenarios with sub-second execution latency requirements; light users who only need simple Zap-style automation at the personal level (Zapier has lower learning costs).

Summary and Outlook of Make

The core value of Make is to integrate the flexibility of visual orchestration and the autonomous decision-making capabilities of AI Agents into the same platform. At the same time, it solves the "black box" problem of enterprise automation through Make Grid and governance modules. It is not the lightest option on the automation track, but at the intersection of the three dimensions of complex business processes, multi-department collaboration and AI-native integration, it forms a clear difference from competing products such as Zapier and n8n.

Current limitations: (1) Self-hosting/privatized deployment is not supported, and organizations with high data sovereignty requirements need to carefully evaluate it; (2) The AI ​​Agent function is still in a rapid iteration period, and the stability of the production environment needs to be verified through PoC; (3) The cost transparency of the credit point billing model for high-frequency execution scenarios needs to be improved; (4) The Chinese interface and document support have not been made public, and the implementation path for the Chinese market is still unclear.

Procurement and Adoption Risk Assessment: It is recommended to use the Free plan to conduct a small-scale pilot on 1-2 high-repetition, low-risk processes (such as form data synchronization, notification automation) to verify whether the construction efficiency and credit consumption are in line with expectations. The pilot cycle is recommended to be 2-4 weeks, focusing on recording: time taken to build a single scenario, daily execution volume, failure rate and recovery cost. After confirming that the model is feasible, consider expanding the Core/Pro solution; before purchasing, enterprises need to confirm the SSO fine-grained granularity, data residency region, and SLA terms in the Enterprise contract. For organizations with an urgent need for private deployment, it is recommended that n8n (open source self-hosting) be evaluated as an alternative.

Related tools: Notion AI, google-workspace

Make’s model and version evolution

Continuous iterative updates, the latest version introduces performance optimization and new features. Historical version information can be viewed on the official release page. There is no complete public version evolution timeline yet. It is recommended to pay attention to the official announcement to understand the rhythm of feature updates.

How to use Make

• Web client: You can use it by visiting the official website and registering an account. Most functions do not require installation.
• API Access: Provides RESTful API, developers can obtain the API Key and integrate it into their own applications.

Version Info

• Make AI Agents ：The AI ​​Agent capability is publicly launched, supporting Agentic Automation, MCP connection and AI application integration (400+ AI integrations), and agent workflow can be built through prompt words or drag and drop. The specific version number and precise release date are subject to the official real-time page. （~2025）
• Make AI capability introduction ：Generative AI nodes (OpenAI, DeepSeek, Perplexity, etc.) are embedded in the platform to support directly calling LLM in the scene for text generation, classification, summary and data extraction. There is no official precise date yet. （~2024）
• Renamed Make ：After being acquired by Celonis, it completed the brand upgrade, reconstructed the interface, strengthened enterprise-level governance capabilities (SSO, audit logs, permission management), and introduced the Make Grid panoramic view to monitor the status of cross-department automation operations. There is no official precise date yet. （~2022-02）
• Integromat ：It starts with visual multi-step scene orchestration, supports branching, aggregation, routing and scheduled triggering, and is positioned as a flexible alternative to Zapier. There is no official precise date yet. （~2015）