Submitting an AI Use Case to Administrative AI Solutions
What AAIS supports
AAIS supports administrative and enterprise units, including finance, HR, facilities, and student services, with AI use cases that draw on institutional data or integrate with campus systems. Examples include a chatbot that answers questions from a unit's document library, or an agent that pulls from an existing system of record.
AAIS does not handle individual productivity use cases. If you want a tool for drafting, summarizing, or working with your own files, see Generative AI Tools Available at UW–Madison (kb.wisc.edu/aisolutions/154501) and the central Generative AI Services page listed below.
Before you submit
Use cases move faster when you arrive with a few things settled:
- A problem statement. What task or decision the AI would support, and who would use it.
- Your data sources. What information the solution would draw on, and which unit is the data steward for it. If your unit is not the steward of the data, identify who is, since their involvement may be needed.
- Data classification. Identify the classification of the data involved. Data classification is the first constraint on any AI use case, ahead of cost or tooling. For privacy, classification, and "OK to use with" guidance, the source of truth is Generative AI Services at UW–Madison (it.wisc.edu/generative-ai-services-uw-madison/).
- A sponsoring unit. The unit that owns the need and would sponsor the work.
How to submit
Submit your use case through the AI use case intake form in Kuali: https://wisc.kualihub.com/app/69e8eeaaf3666f0296efcb20/run. The form captures the problem, data sources, data classification, and sponsoring unit so the review can begin without back-and-forth.
What happens after you submit
- Tracking. You can track the status of your request on the AI Use Case Dashboard.
- Access to the dashboard must be provisioned. Contact the AAIS team to request access.
- Operations Committee review. The AI for Enterprise Operations Committee reviews each use case for feasibility, fit against data classification, alignment with supported platforms, and risk.
- Escalation when needed. Use cases that carry strategic or cross-unit implications are escalated to the AI for Enterprise Steering Committee for a decision.
What to expect on timing
The Operations Committee reviews use cases in quarterly batches. A use case submitted during one quarter is reviewed in the following quarter. For example, a use case submitted in Q3 2026 is reviewed in Q4 2026. Batching the review keeps it on a consistent schedule and gives every submitter a predictable point for follow-up.
Possible outcomes
- Proceed. The use case is a fit and moves into scoping or build.
- More information needed. The committee returns specific questions before deciding.
- Routed elsewhere. The need is real but belongs with another service, and you are pointed to it.
- Not a fit at this time. Often a data classification or feasibility constraint, with the reason stated.
Questions and related resources
- Generative AI Tools Available at UW–Madison (kb.wisc.edu/aisolutions/154501) for individual tools.
- Writing UW–Madison Policies So AI Can Understand Them(kb.wisc.edu/aisolutions/161344) for preparing source content.
- Administrative AI Exchange (kb.wisc.edu/aisolutions/161590) to ask questions and follow updates.
- Generative AI Services at UW–Madison (it.wisc.edu/generative-ai-services-uw-madison/) for policy and data classification.