See the capability. Examine the outputs. Test the difference.
AI has a medical-record problem.
It can't reason from evidence it never received.
Legal AI has become extraordinarily capable at searching, organizing, analyzing, reasoning from and drafting with existing information.
But there is an upstream limitation:
What happens when consequential medical evidence was never developed in the original record?
From evidence retrieval to evidence development.
FEDI™, the proprietary, patent-pending Forensic Evidence Development Intelligence engine within INTELDAS™, is designed to identify consequential medical evidence that should be investigated, quantified or developed beyond what appears in the original record.
INTELDAS then carries that developed intelligence forward into attorney-facing forensic work product.
Traditional AI: What does the record contain?
INTELDAS: What consequential evidence should be investigated that the record does not adequately contain?
Don't read our claim. Examine the difference.
The evidence doesn't start over with each report.
The underlying intelligence carries forward through the claim.
And there is an important distinction:
Thomson Reuters does not have to replace its demand workflow—or use the INTELDAS Demand at all—to test the evidence-development capability.
The INTELDAS Forensic Evidence Audit can operate independently as the medical-evidence layer.
Or the same developed intelligence can continue through the complete INTELDAS architecture.
Click any box below to launch the interactive flipbook viewer preview instantly.
INTELDAS™ did not begin by teaching AI to write a better demand.
It began with a medical-evidence problem.
For more than four decades, Dr. Frank Liberti's work has operated at the intersection of clinical injury evaluation, medical evidence and personal-injury litigation.
More than three decades ago, that work included serving as a medical expert in New Jersey Supreme Court litigation addressing the objective medical evidence necessary to establish qualifying injury under the state's tort threshold—the Verbal Threshold—where the injury claim required proof of significant, consequential and permanent injury.
The lesson was consequential:
The injury alone wasn't enough. The medical evidence had to demonstrate its significance and consequences.
That same problem now exists in a different technological environment.
Today's legal AI can perform extraordinary work with the information it receives.
But AI still cannot reason from consequential medical evidence that was never developed into the case.
INTELDAS grew from more than 13 years of engineering medical-legal forensic methodologies and clinical decision intelligence around that problem.
The result is not simply another demand generator.
It was built to solve the medical-evidence problem before the demand is written.
For years, I kept asking myself a simple question: Why hadn't the legal-AI and demand-formulation companies already solved this?
I believe part of the answer is that we approached the same claim from opposite directions.
That difference matters.
If you begin with the existing record, the natural question is:
“What can AI find, analyze and produce from this evidence?”
My question was different:
“What consequential medical evidence should exist to properly demonstrate this injury—but was never developed in the first place?”
That question ultimately led to INTELDAS™ and FEDI™.
Not simply another way to analyze the evidence.
A proprietary, patent-pending system designed to develop supportable medical-forensic intelligence that wasn't adequately present in the original record.
Perhaps that's why the missing-evidence gap was easier for me to see from the medical side.
I wasn't trying to teach AI to write a better demand.
I was trying to make sure the demand had better evidence to work with.
INTELDAS doesn't ask Thomson Reuters to replace the workflow it is building. It asks what that workflow could do with evidence it never had before.
If developing previously unavailable medical evidence can:
then the value isn't another document.
It's what changes upstream of every document that follows.
And now the proposition becomes even easier to test because the Audit can stand alone.
Thomson Reuters doesn't have to decide whether it wants another demand solution.
It only has to answer one question:
DOES THE EVIDENCE CHANGE?
Take one personal-injury case your team already knows.
Use the same original medical record.
Your existing process does what it does.
INTELDAS does what it does.
Then put the resulting evidence side by side.
If we're wrong, you'll know it.
But if we're right, we should have a very different conversation.