AI use case

A specific scenario in which AI is designed, developed, procured, or used to achieve a particular objective, such as delivering a product or service, enhancing decision making, or providing a defined benefit.

Source: AIEOG AI Lexicon (Feb 2026), adapted from 2024 Guidance for AI Use Case Inventories

An AI use case is a specific application of AI technology to solve a defined business problem or achieve a particular outcome. It describes what the AI is being used for, not how the AI works technically. “Using machine learning to detect fraudulent credit card transactions in real time” is a use case. “A random forest classifier” is a technical implementation detail.

The use case is the unit of governance. When regulators, auditors, and boards ask about AI, they ask in terms of use cases: What is the AI being used for? What decisions does it inform? Who does it affect? What are the risks? Financial institutions should organize their AI governance around use cases because that is how oversight, risk assessment, and accountability are structured.

Examples of AI use cases in financial services:

• Automated transaction monitoring for BSA/AML compliance
• Credit scoring and underwriting decisioning
• Customer identity verification during onboarding
• Document classification and data extraction for regulatory filings
• Chatbot-assisted customer service for account inquiries
• Sanctions screening and name matching
• Predictive analytics for customer churn and cross-selling

Each of these use cases has different risk profiles, data requirements, regulatory implications, and stakeholder impacts. Governing them requires understanding each one individually.

The concept of an AI use case is foundational to AI governance because it provides the organizing principle for risk management, accountability, and transparency. Without clearly defined use cases, institutions struggle to answer basic governance questions: How many AI applications do we have? What are they used for? Who is responsible?

Regulatory expectations are converging around use case documentation. Federal guidance on AI use case inventories, the NIST AI RMF’s emphasis on context and intended use, and examiner questions during supervisory activities all point to the use case as the primary unit of AI governance.

Common challenges:

• Undocumented use cases. AI capabilities can proliferate quickly, especially when embedded in vendor tools. Use cases that are not documented cannot be governed.
• Vague use case definitions. A use case defined as “AI for compliance” is too broad to assess meaningfully. Use cases should be specific enough to identify the data, decisions, and stakeholders involved.
• Missing risk classification. Not all use cases carry the same risk. Institutions need a tiering system that applies governance proportionate to the risk each use case presents.

1. Define use cases with specificity. Each use case should describe the objective, the data inputs, the decisions or outputs the AI produces, and the stakeholders affected.
2. Assign risk tiers. Classify each use case by risk level (high, medium, low) based on factors like regulatory impact, customer impact, decision criticality, and data sensitivity.
3. Assign ownership. Every use case needs a named business owner who is accountable for its governance, performance, and compliance.
4. Connect use cases to your AI inventory. Each use case should link to the specific AI models, systems, and data sources that support it.
5. Review and update regularly. Use cases evolve as business needs change. Conduct periodic reviews to ensure documentation remains accurate.
6. Use the use case as the basis for risk assessment. Each use case should have a corresponding risk assessment that evaluates its specific risks and controls.

AI use case inventory, AI system, AI governance, AI risk assessment, AI lifecycle, Documentation