The Internal Revenue Service (IRS) is investing heavily in artificial intelligence to overhaul its audit selection process, and has awarded Palantir Technologies a $1.8 million contract to refine a custom tool designed to identify high-value tax evasion cases. The move signals a deeper reliance on private technology firms to modernize the agency’s aging infrastructure, despite decades of failed internal attempts.
The Problem: Outdated Systems and Inefficient Case Selection
For years, the IRS has struggled with fragmented data systems, built up over decades, that hinder efficient tax enforcement. According to agency documents, these outdated processes result in duplicated efforts, gaps in coverage, and suboptimal case selection. The IRS currently uses over 100 business systems and 700 methods to identify potential fraud. These efforts have proved increasingly ineffective.
Palantir’s Solution: The “Selection and Analytic Platform” (SNAP)
Palantir’s SNAP software is designed to streamline the identification of fraudulent tax filings. The tool operates as a pilot program, analyzing existing IRS data to flag suspicious cases. The agency has already spent over $200 million on Palantir technology since 2014, and is now seeking to deepen that relationship. SNAP will sit atop the IRS’s splintered databases, helping auditors identify red flags in filings that might otherwise be overlooked.
How SNAP Works: Targeting Specific Tax Areas
The IRS has tasked Palantir with developing three case selection methods:
- Disaster Zone Claims: Auditing relief funds for potential misuse.
- Residential Clean Energy Credits: Scrutinizing tax credits for solar and wind energy installations.
- Form 709 Gift Tax Returns: Examining high-value gifts (artwork, stocks, businesses) for proper disclosure.
Experts suggest SNAP will analyze unstructured data, such as appraisal documents, to verify the accuracy of property valuations. The IRS may also use data from third-party financial platforms like Venmo or Depop, though only if it already possesses that data legally.
The IRS’s History of Failed Modernization
The IRS has struggled for decades to upgrade its technology, with projects repeatedly stalled or abandoned. The agency has even experimented with mining public social media posts and contracting with crypto exchanges like Coinbase to identify unreported income.
Political Obstacles and Staffing Shortfalls
The IRS faces consistent political backlash and underfunding, making modernization efforts difficult. Between February and July 2025, the agency lost over 25,000 employees through resignations and early retirement offers. Frequent commissioner turnover also disrupts long-term projects. Experts note that there is “a lack of political will to die on the hill of the IRS.”
The IRS’s reliance on Palantir reflects a broader trend of government agencies turning to private tech companies to solve modernization challenges. The move raises questions about data privacy, algorithmic bias, and the outsourcing of core governmental functions. The success of SNAP remains to be seen, but the agency’s continued investment suggests it sees AI as a critical tool for enforcing tax compliance.
