Why AI matters at this scale

The Wisconsin Department of Revenue (DOR) is a large state agency responsible for administering tax laws, collecting state revenues, and distributing shared revenues to local governments. With a staff of 1,001-5,000, it processes millions of tax returns, handles complex audits, manages taxpayer services, and provides critical revenue forecasts for the state budget. At this scale, manual processes and legacy systems create inefficiencies, backlogs, and limit the ability to proactively ensure compliance in a complex economic landscape.

For an organization of this size and mission, AI is not a luxury but a strategic necessity to manage operational scale and complexity. The volume of structured and unstructured data—from tax filings and payments to correspondence and call logs—is immense. AI provides the tools to automate routine tasks, derive predictive insights from this data, and reallocate human expertise to high-value, judgment-based activities. This shift is crucial for improving revenue collection rates, enhancing public trust through faster and fairer services, and doing more with constrained public resources.

Concrete AI Opportunities with ROI Framing

1. AI-Driven Audit Optimization: Manual audit selection is time-consuming and can miss sophisticated evasion patterns. Machine learning models can analyze historical audit results, current return data, and external economic indicators to score and rank audit candidates by potential revenue yield. This targeted approach can significantly increase the return on investment for each audit hour, directly boosting state revenues while reducing the burden on compliant taxpayers.

2. Intelligent Document Processing (IDP): The DOR receives a flood of paper and digital documents. IDP using computer vision and natural language processing can automatically extract, validate, and input data from scanned returns, W-2s, and business filings into core systems. This reduces manual data entry errors, accelerates processing times (leading to faster refunds), and lowers operational costs, providing a clear ROI through staff productivity gains and improved citizen satisfaction.

3. Conversational AI for Taxpayer Services: A significant portion of call center and online inquiries are repetitive. Deploying a sophisticated AI chatbot and virtual call agent can handle common questions about deadlines, forms, and refund status 24/7. This deflects a high volume of contacts, reducing wait times, freeing staff for complex consultations, and improving the taxpayer experience—a key metric for public agencies. The ROI is measured in reduced call center costs and improved service levels.

Deployment Risks Specific to This Size Band

For a large government entity, risks are pronounced. Integration with Legacy Systems: The core tax processing systems are often decades-old monolithic applications, making real-time AI integration and data access a major technical hurdle. Data Security and Privacy: Taxpayer data is highly sensitive, requiring AI solutions to meet extreme security standards and strict data governance policies, potentially limiting cloud-based options. Procurement and Vendor Lock-in: The lengthy public procurement process can slow innovation and may lead to dependency on a single large vendor, reducing flexibility. Algorithmic Fairness and Public Scrutiny: Any AI used for enforcement (like audit selection) must be rigorously tested for bias and be explainable to maintain public trust and withstand legislative oversight. A failure here could damage the agency's credibility and lead to legal challenges.