Build AI agents without code – automate with Make and n8n

What will you build in the course?

You will build an enquiry assistant for a fictional company. The assistant receives an enquiry, checks that the required information is available, chooses a category and creates a draft response from an approved knowledge source. The result then goes to a review queue.

The agent may not send the response itself, book an appointment, change an invoice or invent information. The practical exercise therefore shows both what an AI agent can do and how to restrict what it must not do.

• ✓Make: build a visual scenario with a trigger, validation, AI step, router and review queue.
• ✓n8n: build a visual workflow with a trigger, conditions, AI Agent, structured output and safe routing.

You only need to complete one track to finish the course. The other remains available as optional further study.

What you will learn

After the course, you can distinguish between standard automation, a workflow with an AI step and an AI agent. You can also assess whether a task genuinely suits an agent or whether a simpler and cheaper automation is the better choice.

• ✓choose a limited use case with measurable value
• ✓describe the trigger, input, decisions, tools and output
• ✓define allowed and prohibited activities
• ✓build a visual workflow in Make or n8n
• ✓require structured output from the language model
• ✓add human review before external actions
• ✓test normal cases, uncertain enquiries and technical failures
• ✓detect attempts to influence the agent's instructions
• ✓calculate time savings, platform costs and maintenance
• ✓create a 14-day pilot plan

Who is the course for?

The course is suitable for solo business owners, freelancers, employees and small businesses that want to automate recurring tasks without starting with traditional software development. It is also suitable if you have already tried ChatGPT or Claude and want to understand how a response can become part of a reusable workflow.

You do not need to know how to program. You do, however, need to work systematically with instructions, data fields, conditions and tests. No-code means building visually, not that the solution lacks logic, costs or accountability.

What will you take away?

When the course is complete, you will have practical working materials that can be reused:

• ✓a completed agent card
• ✓a workflow map
• ✓copy-ready agent instructions
• ✓a test workflow in Make or n8n
• ✓six documented test cases
• ✓a cost and value calculator
• ✓a 14-day pilot plan
• ✓a monthly quality-assurance routine

The course uses fictional examples so that you can practise without entering real customer data or sensitive information.

What does it cost to use the tools?

Subscriptions to Make, n8n, language models and other external services are not included in the course price. Available features and usage charges can change over time.

The course therefore does not provide a static price list that will quickly become outdated. Instead, you receive a calculator where you enter your current platform, model-run and maintenance costs. This lets you compare the time saved with the full real cost.

The lowest-cost way to start is normally to use a tool you are already allowed to use, work with a small test workflow and retain manual review. Do not buy several tools until a limited use case has passed the test suite.

Make or n8n – which should you choose?

The course does not declare an overall winner. The right choice depends on the systems you need to connect, how the team works, operational requirements and which pricing model suits your usage.

Make provides a visual way to build scenarios with modules and routing. n8n provides visual workflows and can be used as a cloud service or self-hosted. Self-hosting creates more responsibility for installation, updates, security, backups and monitoring. It should therefore not automatically be regarded as free or easier.

In the course, you can build the same type of solution in your chosen tool and assess the result based on quality, cost and your ability to maintain it.

Course format

The course is a self-paced online course of approximately 60–75 minutes. It is divided into short micro-lessons with explanations, interactions, templates, a practical build and knowledge checks.

You can save your responses and continue where you left off. The course ends with a five-question final test, requiring at least four correct answers, and a summary of the material you have created.

Level: specialist course, beginner–intermediate. Prerequisites: no programming required. Tools: choose Make or n8n.

Frequently asked questions

Do I need to know how to program?

No. The practical exercise can be completed visually. You do, however, need to understand how data moves between steps and be able to create simple conditions.

Is n8n really a no-code tool?

n8n describes its platform as low-code because code can be added when needed. The course's n8n exercise requires no custom code and can therefore be completed as a no-code workflow.

Do I need to use both Make and n8n?

No. You choose one main track. The other is optional further study.

Are Make, n8n or an AI model included in the course price?

No. External subscriptions and usage costs are not included. Check the current terms with each provider.

Can I use a free account?

That depends on current plans, the AI features you need and the number of test runs. The course therefore does not promise that the entire exercise can always be completed free of charge.

Will I build a fully autonomous AI agent?

No. You deliberately build a controlled agent that hands its response to a person for review. This is a more suitable starting point for most organisations.

Can the agent send emails automatically?

Such functions may be technically available, but the course prototype sends nothing automatically. You learn to separate an AI-generated suggestion from the decision to use it.

May I use real customer data in the exercise?

No. Use the course's fictional test data. Real use requires the organisation to assess tools, agreements, permissions, data protection and internal rules.

What is the difference between automation and an agent?

A rule-based automation follows a predefined path. A limited agent can choose between allowed tools or steps based on the current task. The course also helps you identify when standard automation is the better choice.

Will I receive a ready-made template?

Yes. The course includes an agent card, core instructions, a test FAQ, six test cases, a value calculator and a pilot plan.

How is the course kept up to date?

Tool features and sources are checked quarterly. Feature and price information is marked with the date it was last checked.