When a company decides to stop moving data manually between systems, the first question that arises is always the same: n8n, Make or Power Automate? All three tools allow you to connect applications and run automated workflows, but their business models, technical limits and implications for data privacy are radically different. Making the wrong choice can mean migrating in twelve months with every workflow rewritten, or paying for licences that cannot be justified by the company's actual volume. This article breaks down the comparison with data updated to 2026 so the decision is technical, not commercial.
What is each tool and who is behind it?
n8n: open-source automation with a self-hosted option
n8n (pronounced «n-eight-n», short for «nodemation») was born in 2019 as an open-source project under the Sustainable Use License. Its key differentiator is that it can be deployed on the company's own infrastructure: a VPS, an on-premise server or Kubernetes. This means that workflow data never leaves the environment controlled by the organisation. In 2025 it launched n8n Cloud (managed SaaS) and surpassed 100,000 starred repositories on GitHub. Its monetisation model is based on the paid Cloud version and on Enterprise plans for self-managed installations that require support and SSO. The active community contributes more than 400 native integration nodes.
Make (formerly Integromat): visual power for complex workflows
Make — rebranded from Integromat in 2022 — is a pure SaaS platform founded in Prague in 2012. Its visual designer in the form of a bubble diagram is particularly intuitive for workflows with conditional branches, loops and complex iterations. Scenario data is processed on Make's servers (currently with EU infrastructure available for paid plans). In October 2020, Celonis — the European leader in process mining — acquired Integromat, accelerating its evolution towards analytical automation and paving the way for the platform's relaunch under the Make brand in 2022. It offers a free plan with 1,000 monthly operations and scalable paid plans billed by operations.
Power Automate: the Microsoft ecosystem
Microsoft Power Automate (formerly Microsoft Flow) is Microsoft 365's automation offering. Launched in 2016, it is deeply integrated with Teams, SharePoint, Outlook, Dynamics 365 and the rest of the Microsoft ecosystem. In 2025 it incorporated autonomous agent capabilities through Copilot Studio, extending its reach beyond linear workflows. Its licensing model is complex: there is a basic layer included in M365 subscriptions and premium plans per user or per flow. Data is processed on Microsoft Azure infrastructure, with EU data residency options for eligible subscribers.
Technical comparison in 2026
| Criterion | n8n | Make | Power Automate |
|---|---|---|---|
| Deployment model | Self-hosted (open-source) or Cloud SaaS | Pure SaaS (Make/Celonis cloud) | Microsoft Azure SaaS |
| Base licence | Sustainable Use License (open-source with commercial restrictions) | Proprietary (SaaS) | Proprietary (Microsoft) |
| Free plan | Yes (self-hosted with no execution limit) | Yes (1,000 operations/month) | Limited (included in M365 with restrictions) |
| Data sovereignty | Full on self-hosted; shared on Cloud | Shared (EU option on paid plans) | Shared (EU residency configurable in Azure) |
| Number of native integrations | 400+ nodes (2026) | 1,000+ applications | 900+ official connectors |
| Learning curve | Medium-high (technical interface, requires node logic) | Medium (intuitive visual, curve on advanced workflows) | Medium (familiar to M365 users, complex on Premium) |
| Custom code | Yes (JavaScript/Python in «Code» nodes) | Limited (built-in functions, no script execution) | Yes (expressions, Azure Functions, custom connectors) |
| Native AI agents | Yes (AI Agent node with configurable LLMs) | Yes (integrated OpenAI and Anthropic modules) | Yes (Copilot Studio, autonomous agents from 2025) |
| Indicative SaaS price (5-user SMB) | From ~€50/month (Cloud Starter) | From ~€10/month (Core, 10,000 ops) | From ~$15 USD/user/month (Premium); basic included in M365 |
| Integration with ERP/CRM | Via REST API or HTTP node; active community | Via REST API or Webhooks; good Salesforce/HubSpot coverage | Native Dynamics 365, SAP connectors; API for others |
| GDPR compliance | High on self-hosted; review DPA on Cloud | Sign DPA; EU infrastructure available on paid plans | Microsoft data processing agreement (GDPR-ready) |
When to choose n8n
n8n is the most suitable option when the company has strict data sovereignty requirements: regulated sectors (healthcare, banking, defence, public administration) or companies that process sensitive personal data and cannot allow that data to pass through third-party servers outside their control. It is also the natural choice when the technical team has the capacity to maintain a self-hosted instance, whether on its own VPS or in an on-premise environment. The open-source code allows every line to be audited, nodes to be tailored to specific needs and removes any dependency on the vendor's roadmap.
Another clear use case for n8n is workflows that require complex business logic written in JavaScript or Python: non-standard data transformations, calls to APIs with custom authentication or integration with legacy systems that have no native connector. n8n's technical flexibility surpasses Make in this area. At Summum we have been building production-grade automations with n8n for over two years for clients in sectors where data cannot leave the corporate perimeter. If you want to explore what n8n can do for your company, take a look at our n8n automation service.
When to choose Make
Make shines when the priority is speed of implementation and the company has no internal technical profile. Its visual scenario interface is the most intuitive of the three for designing workflows with multiple branches, filters and loops without writing a single line of code. It is particularly powerful for marketing automation (connecting CRM + email marketing + social media + spreadsheets) and for teams that need to iterate quickly on their workflows.
Make is also a good choice when operation volumes are predictable and bounded: its operations-based pricing model is transparent and scales reasonably up to certain thresholds. Beyond tens of thousands of daily operations, the cost can become less competitive. It is worth noting that, as a pure SaaS platform, the company must sign a Data Processing Agreement (DPA) with Make/Celonis if it processes personal data in its workflows, and must explicitly select European infrastructure to comply with the GDPR.
When to choose Power Automate
Power Automate is the obvious answer when the company is already in the Microsoft 365 ecosystem. If employees work in Teams, SharePoint and Outlook, Power Automate reduces integration friction to zero: native connectors are certified by Microsoft, flows are managed from the same M365 admin portal and basic licences are already included in many subscriptions. For automating document approvals, calendar notifications or SharePoint list synchronisation, Power Automate wins on ease of use at no additional cost.
Complexity arises when connectors are needed for applications outside the Microsoft world: Premium licences (per user or per flow) are required at a price that can be steep for SMBs with many users. Power Automate is also not the best option for workflows that require advanced programming logic or full data sovereignty. Its Copilot Studio agent ecosystem is promising but adds another layer of complexity and cost. For companies with hybrid environments (M365 + proprietary systems), our enterprise automation practice can help you decide whether Power Automate covers the use case or whether it needs to be complemented with another orchestration layer.
The dimension that never appears in demos: GDPR and data privacy
One of the most common mistakes when deploying an automation tool is failing to ask what data flows through the workflows. When a Make scenario or a Power Automate flow moves information about customers, employees or patients between systems, that tool becomes a data processor under Article 28 of the General Data Protection Regulation (GDPR, EU Regulation 2016/679). This implies:
- Signing a DPA (Data Processing Agreement) with the provider.
- Verifying that the provider does not transfer data internationally to third countries without adequate safeguards (Adequacy Decision or Standard Contractual Clauses).
- Recording the workflow in the company's Record of Processing Activities.
- Assessing whether the automation requires a Data Protection Impact Assessment (DPIA) if it processes data at scale or involving special categories.
n8n self-hosted eliminates the root cause of most of these issues because data never leaves the company's server. Make and Power Automate handle them reasonably well, but require explicit configuration and documentary review. If your company operates in a regulated sector or has doubts about the legal basis of its automations, it is worth reviewing the compliance framework before deploying any tool.
Real use-case comparison
| Use case | Best option | Reason |
|---|---|---|
| Synchronise CRM with ERP in a company handling sensitive data | n8n self-hosted | Data sovereignty, no data leaving the perimeter |
| Automate multichannel marketing campaigns (CRM + email + Slack) | Make | Fast visual interface, broad coverage of marketing SaaS apps |
| Invoice approval workflow in a company using M365 | Power Automate | Native integration with Outlook, Teams and SharePoint; licence already included |
| AI agent that queries an internal database and generates reports | n8n | AI Agent node with configurable LLM, execution on own server |
| Quick integration of SaaS tools (Notion + Google Sheets + Slack) | Make | Native connectors, low learning curve, competitive price at low volume |
| HR process automation in a company using Dynamics 365 | Power Automate | Native Dynamics connector, centralised management in M365 admin |
| Workflows requiring complex data transformations or custom API calls | n8n | JavaScript/Python code nodes, maximum flexibility |
Can they be combined? The multi-tool approach
In mid-size companies with heterogeneous technology environments, a combined approach is common: Power Automate for internal M365 workflows, n8n for automations that touch proprietary systems or sensitive data, and Make for quick integrations with marketing or sales SaaS apps. This approach is not chaotic as long as a clear design criterion exists: each tool has its own responsibility lane and the workflow map is documented so the technical team can maintain it.
The risk of the multi-tool approach is fragmentation: duplicated workflows, data inconsistencies and licence dependencies on three different providers. That is why we recommend starting with a single tool (usually n8n or Make depending on the company's profile) and expanding only when a specific use case justifies it.
Frequently asked questions
Is n8n really free?
The self-hosted version of n8n is open-source and can be deployed without a licence fee on your own server. However, it has indirect costs: infrastructure (a basic VPS can cost between €5 and €20 per month depending on the provider), technical maintenance and updates. For companies without an internal technical team, the paid Cloud version simplifies operations but carries a monthly cost. n8n's Sustainable Use License prohibits offering n8n as a managed service to third parties without a commercial agreement with n8n GmbH, but there are no restrictions on internal business use.
What happens to my customers' data if I use Make or Power Automate?
When workflows process personal data, both platforms act as data processors under the GDPR. Make offers European infrastructure on its paid plans and facilitates signing the DPA. Microsoft Power Automate processes data in Azure with a dedicated data processing agreement for European customers. In both cases, it is the responsibility of the data controller (your company) to verify that the provider meets the conditions of Article 28 of the GDPR and to record the processing relationship in the Record of Processing Activities.
How long does it take to deploy a basic automation?
A simple automation (for example, «when a contact is created in the CRM, add them to a mailing list and notify via Slack») can be configured in under an hour with any of the three tools. Complex workflows with conditional logic, error handling and retries may require between one and three days of technical work. The real difference appears in maintenance: workflows grow over time and without documentation or version control they become technical debt. Getting it right from the start — with clear naming, error alerts and testing — saves weeks of work in the following months.
Does Power Automate include desktop automation (RPA)?
Yes. Power Automate includes Power Automate Desktop, a robotic process automation (RPA) tool that can interact with desktop applications, websites and legacy systems without an API. This capability is unique compared to n8n and Make in their native versions, although n8n can be complemented with external RPA tools. Power Automate Desktop is included for free in Windows 10 and 11, although unattended executions (without an active user) require additional licences. For SMBs with manual processes in applications that have no API, this feature can be decisive.