Syllabus

Title
2581 From Models to Agents: Agentic AI in Business I
Instructors
Dr. Daniel Ringel
Contact details
Type
PI
Weekly hours
2
Language of instruction
Englisch
Registration
10/01/26 to 10/30/26
Registration via LPIS
Notes to the course
Dates
Day Date Time Room
Wednesday 11/18/26 01:00 PM - 05:00 PM TC.5.14
Wednesday 11/25/26 01:00 PM - 05:00 PM TC.4.17
Wednesday 12/02/26 01:00 PM - 05:00 PM LC.5.096
Wednesday 12/09/26 01:00 PM - 05:00 PM TC.3.12
Wednesday 12/16/26 01:00 PM - 05:00 PM D3.0.237
Contents

This hands-on doctoral course introduces effective use of contemporary AI for business  analytics and research. Students learn to specify and verify model assisted work, build a  small analytics application through AI assisted coding, call a model through an API, obtain structured output, and connect the model to controlled analytical tools. The course  culminates in a minimal bounded agentic AI assistant. The emphasis is on durable principles, analytical correctness, and verification.

SessionCentral questionPrincipal output
1How do I work effectively with an AI model?AI task specification and verification record.
2How do I build software through AI assisted coding?Working analytics application and failure log.
3How does an application communicate with an AI model?Reproducible model measurement pipeline.
4How can AI perform a bounded analytical workflow?Model and tools analytical application.
5When does an AI application become an agent?Minimal bounded agent and capstonedefense.
Learning outcomes

After completing the course, students can (1) use AI through clear task specifications and  staged interaction; (2) verify AI supported analytical work; (3) build and repair a small application with an AI coding assistant; (4) make basic model API calls and process structured output; (5) connect a model to controlled analytical functions; (6) distinguish  models, applications, workflows, and bounded agents; and (7) document and defend AI assisted work reproducibly.

Attendance requirements

Attendance and active participation are mandatory because the course consists of five cumulative sessions. Students participate in the guided labs, peer testing, failure debriefs,  and final project defense. Absences are handled according to the applicable WU study  regulations.

Teaching/learning method(s)

Each session begins with a 30-minute discussion of one academic pre-reading. A concise conceptual input and reference demonstration are followed by a guided hands-on lab, an independent challenge, and a failure debrief. All sessions use supplied starter repository and/or code base with data. AI assisted coding is an explicit learning objective.

Assessment

Individual AI analytics competency task (20 percent); AI assisted analytical application with  tests, process documentation, and automated acceptance checks (35 percent); models to agent project with evaluation, report, live demonstration, and individual defense (45 percent). All three components must be completed.

Grade | % Max Points
1 | > 89
2 | 80 - 89
3 | 70 - 79
4 | 60 - 69
5 | < 60
Readings

Please log in with your WU account to use all functionalities of read!t. For off-campus access to our licensed electronic resources, remember to activate your VPN connection connection. In case you encounter any technical problems or have questions regarding read!t, please feel free to contact the library at readinglists@wu.ac.at.

Recommended previous knowledge and skills

No previous API experience is required. Prior Python experience is welcome but not  necessary. Students complete the technical setup and diagnostic task before Session 1 and bring a laptop capable of running a Python environment and installing Claude Desktop. 

Availability of lecturer(s)

Before and after class and by appointment. Questions that are relevant to the full group should be submitted through the designated course channel. Research design questions may  also be addressed in scheduled project clinics.

Other

The course is taught in English. Students use supplied repositories and codebases as well as public or instructor supplied data. The reference software environment and model versions are pinned for the course. Alternative tools are permitted, but support is limited to the reference stack. Course managed access and usage limits are communicated before the first session. Tool access, hosting, and data protection follow the separate Academic 
Dependencies and Data Protection Note. Students are expected to use AI-supported coding for the duration of the course (i.e., Claude  Code with a Claude Pro subscription) and have (or obtain) an Anthropic developer account with sufficient  API credits (approximately EUR 50).

Last edited: 2026-09-25



Back