Class Overview
What will we do?
This class will serve as our primary introduction to the concept of agentic automation, which involves the use of AI agents that can perform tasks autonomously and adaptively. We will discuss how agentic automation differs from traditional RPA and explore its potential applications in accounting and finance. The class will include a case focused on variance analysis, a common task in accounting and finance that involves analyzing the differences between planned and actual performance. We will first review how this process is typically done manually, and then discuss how it can be enhanced using a combination of RPA and agentic automation. The RPA part of the solution will be demonstrated using UiPath's platform, while the agentic part will involve using a large language model (LLM) to assist with identifying the sources of variance and providing insights for decision-making. This class will provide students with a practical understanding of how to design and implement intelligent automation solutions that combine both RPA and AI agents and important step toward designing agentic workflows.
Materials and Preparation
Class Materials
- Link: Class 5 Template
- Link: Class 5 Files (*.xlsx in a zip folder)
- Slides: PowerPoint or PDF
- Automation Tools: RPA Demo: UiPath Web Studio (cloud)., RPA with Agent Demo: UiPath Maestro (cloud).
-
Suggested Pre-Class Preparation
There is no suggested preparation for this class. -
Class Plan
- We will start with a brief review of financial reporting material from earlier classes and contrast it to how managerial accounting is used across the accounting and finance function.
- This review will include a demonstration of a workflow for an external user of accounting information (e.g., an investor) undertaking an analysis of the revenue of a company, separating out the role of RPA from the role of an agent.
- Our content focus area for this class is variance analysis. We will undertake a variance analysis case together as a class to understand how this process works.
- After a review of the manual process, we will discuss how it can be improved using a combination of automation tools including separating RPA steps from agentic steps.
- The RPA part of the solution will be demonstrated as an automation solution using UiPath's platform for our small simulated case data.
- Students will be asked to reflect on how this process could scale and how they would modify the RPA solution to work at a larger scale.
- The final part of the case has students investigate how an LLM, acting as an agent, could potentially help with pain points relating to tracking down the sources of budget variance and aid their decision-making.
- After completing the case, there will be time to work on the final team project, each team is expected to check in with me regarding the area in which the team will focus on from the list of focus areas, or propose an alternative question they will solve.
Required Deliverables
| Deliverable | Due Date | Canvas Submission Portal |
|---|---|---|
| Professionalism (individual): Agentic Automation Submission | During Class | Upload to Canvas |