Workshop: From AI chat to practical AI agents and better business decisions.
Build three practical AI agents using your own meetings, documents and data. Leave with working prototypes and a 90-day AI action plan.
What would you hand over to an AI agent if you could free up time for more important decisions?
On 12 November, Lifestyle and Design Cluster and the University of Southern Denmark (SDU) invite CEOs and their colleagues from the fashion, furniture and design industries to a six-hour, hands-on workshop.
The goal is to move beyond experimenting with AI chat tools and start building AI into the way your company works.
Throughout the day, you will work with your own company documents, data and business challenges. You will build two AI prototypes, explore how agents can support management decisions, and develop a practical plan for the first 90 days.
No developer is needed. Everything you build is yours to keep.
This is a working session, not a traditional conference. You will leave with:
The focus is not on automating everything. It is on understanding where AI can create business value, where a simple prompt is enough, where an agent makes sense, and which decisions must remain with people.
09:00 – Welcome: How AI agents actually work
We start with one question: What is the first task you would hand over to an AI agent?
You will be introduced to the four essential parts of an agent: a source of information, a structure, an instruction and a human checkpoint.
A live demonstration shows how a real meeting transcript can be turned into decisions, responsibilities, deadlines and a follow-up email. You will receive the templates to use in your own organisation.
09:50 – The Agentic Company
Oliver presents his new paper on the agentic company, followed by questions and discussion.
10:15 – Build 1: Your company knowledge base and the agent that reads it
Where should company knowledge live, how should it be structured, and what information should an AI agent never access?
We compare how an AI agent performs with poorly organised files versus a structured knowledge base.
You will then build and test a question-answering agent using a small selection of your own company documents, including testing questions it should refuse to answer.
The same approach can support HR handbooks, supplier and certification tracking, onboarding and internal support.
You leave with a working agent and a practical folder and naming standard.
11:15 – Break
11:30 – Meet the companies: Connecting business and talent
Students from SDU join the workshop.
Each company briefly introduces its business, its AI ambitions and the skills and profiles it may need in the future.
The conversation continues over lunch, with each company hosting its own table.
12:00 – Lunch with the students
12:45 – Build 2: From business data to a decision brief
Bring one Excel or CSV export, such as sales by style and account, stock coverage or another recurring business report.
You will build an agent that checks the numbers against rules you define and produces a short management brief:
You define the thresholds. The agent identifies what needs attention. A person makes the decision.
The same approach can be used for board reporting, strategic priorities and sustainability targets.
You leave with your own rule sheet and a first decision brief based on your data.
13:50 – What AI agents cannot do, and what the law requires
AI agents can produce confident but incorrect answers, lose important context or fail to recognise the human consequences of decisions.
We explore where human oversight is essential, particularly when decisions affect people, finances, safety and reputation.
We also discuss the EU AI Act in practical terms, including the distinction between lower-risk administrative support and high-risk applications such as recruitment screening.
You receive a two-page AI policy template that can be adapted for your company.
14:30 – Your AI setup and the first 90 days
What should your company's AI operating system look like?
We map potential applications, from personal assistants and project management to strategy support and management reporting.
You will identify which tasks deserve a dedicated AI agent and which can be handled with a well-designed prompt.
Before leaving, you will select one solution to implement within 30 days, define who owns it, agree on what success looks like and establish the basic rules for data use and human approval.
15:00 – End of workshop
The workshop is designed for CEOs and leadership teams who have already started exploring AI and want to move towards practical business applications.
We encourage each company to bring the CEO and one or two relevant colleagues, connecting management responsibility with operational knowledge and implementation.
You do not need programming skills or previous experience building AI agents.
You should, however, be ready to work with your own business challenges and actively participate in building the solutions.
We will work with your material throughout the day. To make the hands-on sessions valuable, please bring a laptop and prepare the following:
1. Company documents for your knowledge agent
Bring 10–20 recent documents, such as employee handbooks, policies, project briefs, meeting notes or internal guides.
Files should be in Word format or text-selectable PDFs. Scanned documents and photographs of paper documents will not work.
2. Business data for your decision agent
Prepare one Excel or CSV export covering approximately 6–12 months.
Examples include sales by product and customer account, stock coverage or a report you already receive regularly.
The file should contain one clear table, consistent dates and no merged cells or pivot tables.
Also prepare 3–5 simple business rules you would like the agent to apply. For example: "Flag any style with more than 16 weeks of stock."
3. Optional meeting transcript
Bring a transcript from a recent leadership or project meeting, ideally 30–60 minutes long.
A Teams or Zoom text export, Word document or audio recording can be used. A sample will be available if your own material is too confidential.
Important: Remove personal information about customers, employees and job candidates, as well as sensitive or confidential information that should not be shared. Please send your data export one week before the workshop so the instructors can check that it can be used.
Date: 12 November 2026
Time: 09:00–15:00
Location: University of Southern Denmark (SDU), Odense
Language: English
Organised by: Lifestyle and Design Cluster in collaboration with SDU
Lifestyle and Design Cluster members: DKK 2,500
Non-members: DKK 4,500
The workshop has very limited capacity to allow time for hands-on work with each participating company's own material.
Bring your leadership team, your business challenges and your data. Leave with working AI prototypes and a practical plan for what comes next.
