Design of AI-agenter (5-6 May, online, english)

2 days, in english

AI agents amplify your work capacity by handling repetitive tasks — so you can focus on the creative and strategic work that truly creates value. This course teaches you to build AI agents that raise quality while eliminating the boring parts of the job.

During the course you work hands-on in Abundly’s agent platform, but the principles and design methods you learn are applicable regardless of which platform you use in the future.

Target Audience

This course is for you if you already use generative AI in your work and want to take the next step — building your own AI agents that actually work in practice.

The course is relevant regardless of what function you work in. Whether you’re in finance, marketing, HR, operations, product development, or any other area — the content can be adapted to your daily work and workflows.

You’re a great fit for this course if you:

Want to go from using AI in a chat to building and using AI agents
Have concrete workflows you want to automate or improve
Want to understand how to design agents that deliver real value

After this course you can…

✅ Design effective AI agents with the right prompting techniques and LLM models
✅ Build working agent prototypes for your own workflows
✅ Identify processes with the highest potential for AI agent implementation
✅ Understand security aspects and protection mechanisms for agents
✅ Test and quality-assure agents systematically using evals
✅ Integrate agents with external systems via APIs and MCP
✅ Optimize agent costs through the right choice of model and architecture
✅ Build interactive apps and dashboards with agents
✅ Compare and choose between different agent platforms

Course Content – Overview

Fundamental Understanding — AI agents, context engineering, prompting, and LLM models
Agent Design & Implementation — Agent Design Canvas, tools, function calling, and RAG
Practical Exercises — Build your own agents for your workflows
Interactive Apps — Dashboards, forms, and React apps
Security — Prompt injection, protection mechanisms, and security layers
From Canvas to Prototype — MVP thinking and prototype building
API Integrations — HTTP requests and MCP
Evals — Systematic testing and validation of agent responses
Content Generation — Text generation, image generation, cultural adaptation, and delegate task
Cost and Performance Optimization — Model selection, tokens, and optimization strategies
The Agent Platform Landscape — Comparison of tools and platforms
Next Steps — From prototype to pilot to implementation

Teaching Methods and Course Structure

The course combines theory with extensive practical exercises where you create agents for your own workflows. Through demonstrations of real use cases such as news agents, invoice routing, and screening agents, you get concrete examples to build upon. You work in small groups to solve problems and gain hands-on experience with Abundly’s agent platform throughout the course.Get feedback and iterate on your prototypes

Tools & Platform

As our course platform we use Abundly’s platform — an agent platform that cherry-picks the best from leading AI tools. Claude for reasoning, Gemini for image generation, ChatGPT for speech-to-text, Perplexity for research — all in one interface. You focus on what you want to accomplish, not on managing tools.

We also use other AI tools during the course — which ones depends on what participants have access to and what is currently best suited for the task.

What you build during the course is not locked to Abundly — agent instructions and principles can be transferred to other platforms.

You get access to the platform during and after the course.

The course is relevant whether your organization currently builds agents on Claude Cowork, Claude Code, Microsoft Copilot Studio, Google AgentSpace, n8n, or any other platform.

Voices from Previous Participants

What’s Included

Course materials and exercises
Access to Abundly’s agent platform (app.abundly.ai) during and after the course
Course certificate

Prerequisites

Good experience using AI tools like ChatGPT or Claude

Course Content – More in Detail

Fundamental Understanding

What is an AI agent and how does it differ from traditional AI solutions?
Basic prompting and refinement
Different types of LLMs and when they should be used
The difference between autonomous, semi-autonomous agents, and assistants

Agent Design & Implementation

Our Agent Design Canvas — a framework for structuring agent projects
Tools and function calling
Context management — how to give your agent the right information
Practical exercises: Build agents for your own workflows

Security and Quality

Agent security and protection mechanisms
Hallucinations: how to handle and minimize incorrect information
Validation of agent understanding

Practical Applications

Demonstrations of real-world use cases
Process mapping and identifying optimal areas for agent implementation
Create agents relevant to your work and your organization

Agents in the Organization

How to work with agents in teams
Feedback and improvement cycles for agents
The way forward — how to continue working with agents after the course
From prototype to pilot to implementation

Evals and Quality Assurance

Systematic testing of agent responses with evals
Creating and refining graders
Comparing results between different models
Iterating instructions based on eval results

Content Generation and Creative Work

Image generation with AI
Culturally adapted text generation
Reflection prompts and thinking tokens
Sub-agents and parallel processing (delegate task)

API Integrations and Platform Selection

HTTP requests to external APIs
Model Context Protocol (MCP)
Overview of the agent platform landscape

Cost and Performance Optimization

Tokens, model sizes, and pricing
The strategy ”Make it work, make it safe, make it cheap”
Delegation to cheaper models with sub-agents

Career Opportunities and Demand

AI agent expertise is becoming increasingly sought after in the job market as companies seek ways to automate complex processes and improve efficiency. Competence in agent design and implementation opens doors to new roles and career opportunities.

AI Agent Specialist
Process Developer with AI focus
AI Architect
Automation Engineer
AI Consultant with agent expertise

Next Steps

For Your Organization

The course is often the starting point for a larger AI transformation. We help organizations move from individual agent experiments to systematic change:

Start with a pilot — In 4–6 weeks, we help you build a first value-creating agent while training your team. You get hands-on support, see concrete results, and understand how human+AI teams can transform your work. Perfect for testing and seeing the value of AI agents with us as your partner.

Scale through close collaboration — For organizations that want to transform broadly. You get platform access, training programs for all levels, co-development of agents, and ongoing coaching. We become your dedicated partner in building human+AI capability throughout your organization.

Our philosophy is augmentation, not just automation — AI agents that amplify human capacity, not replace it.