Data Trace Report on AI Search of Public Records

Data Trace Information Services LLC, released a new analysis “AI Title Search Tested in the Real World: What Accuracy, Risk, and Readiness Really Look Like,” examining whether public-record-only AI can deliver the completeness, confidence and insurability required to support insurable title decisions at scale. Data Trace framed the question as whether AI operating on public records alone can consistently produce the complete, validated information required to support insurable title decisions.

The findings reinforce how often meaningful misses occur and why those misses matter for insurability. Among the paper’s key findings:

• In a review of 200 residential title files, public-record-only AI search missed at least one meaningful matter in 40.8% of searchable files. When compared with title search supported by DataTrace plant data, these misses highlighted significant gaps in accuracy.
  
  
• The most significant gaps occurred in high-risk categories, such as involuntary liens, which had an issue fail rate of more than 36%. 
  ​
  
• Of the 200 residential title files evaluated, AI was unable to search 16 files (8%) because they lacked the title plant data or comparable normalized datasets needed to complete the search, highlighting an important limitation of public-record-only AI title search.