In Anderson, many cases involve facts that are hard to model generically—such as visibility issues, speed and braking distances on local roads, roadway design, or how quickly emergency response and documentation were created after the incident. AI tools typically ask for basic details and then output a range. The problem is that two cases can have the same “type” of incident and still produce very different outcomes depending on what can be proven.
Common reasons AI estimates can be off include:
- Unclear fault (multiple actors, disputed causation, or conflicting witness accounts)
- Missing documentation (photos, scene notes, vehicle data, medical records)
- Causation disputes (defense arguments that the death was caused by something other than the wrongful act)
- Coverage and policy limits that affect what insurers will actually pay
If you’re relying on an online estimate while the case is still developing, it can become an anchor—one that’s difficult to undo later.


