n8n
“Automation / AI integration” — an automation engineer you bring in-house
What is n8n?
n8n is a workflow automation tool from Germany-based n8n GmbH. You can link apps by drawing a flowchart, or install it on your own company server and add AI agents without sending data outside your company.
- 1Install it on your own servers, and your data never leaves the company
- 2The AI agent node is built in by default
- 3The self-hosted Community Edition is free with no time limit
What sets it apart
These are the differences we found by running the same material through several tools in our training sessions.
You can keep it on your own server
You can run it on your company server or an in-house computer. The official docs state that the self-hosted Community Edition is free with no time limit and includes nearly all features. It is a practical option for organizations that shelved automation because data could not leave the company.
AI agent built in by default
Add an AI Agent node to a workflow and, instead of following a fixed sequence, it chooses the right tools for the goal on its own and loops as many times as needed to finish the job. AI isn’t bolted onto an automation tool here; the product is built around it.
When you get stuck, you can fix it yourself
Even without a prebuilt connector, you can call an API endpoint directly or drop in a short piece of code. For services not on the list, such as Korean ones, being able to connect them today instead of waiting makes a big difference.
⚖️ The weaknesses are clear, too
Someone has to own setup, security and backups, and terms like nodes and credentials feel like a wall at first. It is the hardest of these tools for non-developers to start on their own, and with self-hosting, when problems arise you dig through docs and community forums instead of calling a help desk.
💡 In plain EnglishInstead of outsourcing the work to an agency, you bring an automation engineer in-house. It takes more effort, but you hold the keys.
In organizations with strict security reviews, automation gets blocked not because of missing features but because of one line: 'data passes through an external server.' Being able to install and run it yourself removes that line.
Same job, different approach
Security review has rejected automation twice
The old way
- Build a pilot workflow with an external automation service
- Security review flags that 'data leaves the company'
- Staff end up copying data over by hand again
- Repeat the same discussion from scratch next quarter
With n8n
- Install it on an internal server and build a pilot workflow
- Shows in a diagram how your data stays inside the company
- Flag what leaves your environment at the AI steps only
- The review scope narrows from the whole tool to a single AI step
Common uses
Organizing materials subject to export controls
Move and cross-check data that is hard to send outside, such as HR, payroll and customer lists, entirely inside an n8n instance installed on your own servers.
Inquiry triage agent
An AI agent reads incoming inquiries and sorts them by type. For common questions it drafts a reply, and anything ambiguous goes to a person.
Connecting services not on the list
For targets with no off-the-shelf connector, such as internal systems or Korean services, you call them directly by URL. You don't have to wait for a connector to appear.
Who it’s for
Best for
- ✓Organizations that have put off automation because of data-export restrictions
- ✓You want to put AI agents inside your business workflows
- ✓Organizations with in-house staff to manage servers
Better served by other tools
- –Anyone who wants to build a single workflow on their own today
- –Organizations with no one to handle installation and backups
- –You need a Korean-language interface and phone support
Can we use it at work?
The questions that adoption leads and security teams ask first.
Note: the license for the self-hosted Community edition differs from typical open-source terms. Before adopting it in-house, read the official license text and check where the enterprise features begin.
10-minute quick start
- 1
First, pick a deployment: Cloud if you want no maintenance, or self-hosted if you want to keep your data in-house.
- 2
For the cloud version, sign up at n8n.io and start right away. For self-hosting, follow the self-hosting guide in the official docs (Docker or npm).
- 3
Keep your first workflow very small, such as receiving data through a webhook and writing it to a sheet.
- 4
Once you’re comfortable, add an AI Agent node and attach the workflows you built earlier as that agent’s tools.
Copy-and-paste prompts
Replace only the [bracketed] parts, then paste it in.
Pass the inquiry text received via webhook to an AI agent. The agent should classify the inquiry into one of four types: 'quote, schedule, complaint, other'. For a quote, find the matching course in [course list] and note an estimated range. For a complaint, hand it straight to a person. Also leave the reasoning for its decision in one line.
In a flow installed on our internal server, when an employee question comes in, answer using only evidence found in the internal policy documents. If the documents have no basis for an answer, don't make anything up. Reply 'Not specified in the policy' and point to the responsible department.
Every day at 9 a.m., send a request to [Korean service URL] and fetch yesterday's submissions. Compare the list with the Google Sheet [intake log], add only the entries that are missing, and if the request fails, retry up to twice at 3-minute intervals.
When a PDF is uploaded to the Drive folder, split the document and summarize it, then save the summary to our internal database. If many files arrive at once, queue them and process them one at a time. Send the admin channel a status update once a day.
Every day at 2 a.m., collect the executions that failed in the past 24 hours and group them by cause. If the same cause appears 3 or more times, notify the owner; otherwise, just log it. In the report, include each failed workflow’s name and its last error message.
Using it with other tools
Linking two or three tools changes your work more than any single tool does.
Internal document Q&A
Intake → Alert → Report
Before you start
- Without backups, a self-hosted install loses all your workflows the day the server fails.
- The moment you attach an AI node, that step's content leaves your environment. Self-hosting is no reason to relax.
- If credentials are stored under a personal account, workflows stop when that person leaves the role.
- Node behavior can change when you upgrade versions. Upgrade in a test environment first.
Pricing: personal use and work use
Plans that individuals pay for are different from plans that organizations contract for by member.
For individual use
Run it on your own server. According to the official docs, there is no time limit on use
2,500 executions per month · 5 concurrent executions · 2,300 AI credits per month
10,000 executions per month · 20 concurrent executions
Using it at work
40,000 executions per month · self-hosting available · single sign-on (SSO, SAML, LDAP)
Priced by execution count · n8n-hosted or self-hosted
Checked 21 Sep 2026 · Amounts are shown in euros on the official pricing page and assume annual billing. Raising the execution count also raises the price within the same plan. Prices can change without notice, so check the official page before you buy. n8n official pricing ↗ · n8n official docs: choosing how to use n8n ↗
Agentic AI — the version that does the work itself
n8n's AI features are not bolted-on decoration. They are built in as Agent nodes that make their own decisions within the flow and loop multiple times.
AI Agent node
The official docs describe it this way: 'the agent uses a language model to decide for itself which actions to take.' Unlike a chain that follows a fixed sequence, it picks the tools that fit the question and **runs several times**, splitting the work into preparation, tool calls, result evaluation, and the answer. You can also attach other workflows you have built as its tools.
- No installation needed: sign up for n8n Cloud and add the 'AI Agent' node to your workflow.
- Install it yourself — follow the self-hosting guide in the official docs (Docker or npm) to put it on your server, then use the same nodes.
- Connect a model — register the key from your AI model provider as a credential.
Installation methods and conditions follow the official docs. n8n Docs — What Agents Do ↗ · n8n Official Pricing ↗
Where n8n falls short
We don’t just list the positives. Here is what fell short when we used it.
- Big jumps between pricing plans
On the cloud plans, the step above Pro at €50 per month is Business at €667 per month, with nothing in between. Your team grows only slightly, yet the next option jumps sharply, and every time you go a little over your execution count you have to recalculate.
- Korean-language resources are hard to find
The interface and the official docs are in English, so when you get stuck you end up translating overseas community posts to follow along. There are still few case studies showing how Korean business processes were connected, so even finding a similar workflow takes a long time.
- It becomes a workflow only its creator understands
Once a workflow passes about twenty nodes, it’s hard to take in at a glance. If you don’t annotate each step, you have to re-read even your own workflow later, so you end up writing a separate handover document.
Tools to compare
MakeAutomation built like a mind map→
ZapierThe world’s largest workflow automation tool→
SolapiText and Kakao messaging via API→FAQ
Last updated 22 Sep 2026 · The ratings on this page come from John Kim, CEO of Nanum Consulting, who has used these tools firsthand in training and real work.
n8n, Powered by Your Own Team
Send us a message and we’ll reply within 24 hours. We’ll propose a curriculum tailored to your industry and job roles.
Courses that teach this toolAI Tools Training (by Use Case)