- AI evaluations in indirect tax often look strong in demos but fail in real-world use because they don’t test compliance risk, auditability, or jurisdictional complexity.
- A major red flag is when AI outputs are confident but not traceable or auditable, making it hard to defend determinations to auditors.
- Another red flag is when performance is shown only under ideal, controlled conditions instead of messy real data with errors, gaps, and changing rules.
- Broader warning signs include claims of full autonomy, use of generic rather than tax-specific training, poor error-handling visibility, and high-level promises that don’t hold up under scrutiny.
Source: tax.thomsonreuters.com
Note that this post was (partially) written with the help of AI. It is always useful to review the original source material, and where needed to obtain (local) advice from a specialist.
