Tutorial ID | T-28 |
Section | Reporting & Billing |
Title | Adding your own conclusions and expert opinion to the report |
Subtitle | Add expert opinion that's clearly labeled, consistent with the AI content, and readable by a non-medical audience. |
Before You Begin
You have a complete AI-generated report for the case, with all workflows finished running.
You know who will read the report and whether they have a clinical background.
What You Will Accomplish
By the end of this tutorial, every conclusion you add will be clearly separated from the AI-generated content, framed as professional opinion rather than fact, consistent with the rest of the report, and written in language the intended reader can follow.
Why Does this Tutorial Matter
A report that combines AI-generated content with an expert's own conclusions only works if the reader can tell which is which, trust that the two agree with each other, and understand what each conclusion actually claims. Getting this right is what gives your professional judgment its full weight in the report.
Get it wrong, and a reader can't tell whether a statement reflects a document or your own judgment, a conclusion written as fact rather than opinion becomes vulnerable if it's later disputed, a conclusion that contradicts the AI-generated content makes the whole report look unreliable, and a conclusion drawn before you've read the full report — or written in language the reader can't follow — may need to be revised after it's already been shared.
What to Do if Something Goes Wrong
The list below covers the most common problems when adding your own conclusions to a report. Each entry follows the same pattern: what you'll notice, why it likely happened, a real example, and how to fix it.
Problem: A reader can't tell whether a statement in the report reflects the AI's processing of the records or the expert's own judgment.
Likely cause: Expert conclusions were added without a clear structural or visual separation from the AI-generated sections around them.
Example: A sentence expressing the reviewer's opinion on causation sits inline within an AI-generated paragraph, with nothing to mark where one ends and the other begins.
Fix:
Before adding conclusions, identify the sections where expert opinion will appear and confirm they're clearly labeled and structurally distinct from the AI-generated sections.
Write each conclusion in its designated section rather than inserting it inline within an AI-generated paragraph.
Review the report and read through any section where AI-generated content and expert opinion sit close together, confirming the boundary is clear to a first-time reader.
If the distinction isn't obvious, add a heading or introductory sentence that explicitly identifies the following content as the reviewer's professional opinion.
Use the Comment function to separate your own conclusions from AI-generated content by highlighting the relevant text and adding a comment.
Problem: A conclusion about causation, severity, prognosis, or standard of care reads as an established fact rather than a professional judgment.
Likely cause: The conclusion was written in the same declarative style as a factual statement, without any framing that identifies it as an interpretation.
Example: A sentence states “The injury was caused by the fall” with no framing, when the records support that conclusion but don't establish it outright.
Fix:
Read through the report and flag every statement that makes a claim about causation, severity, prognosis, standard of care, or consistency with an injury mechanism.
For each flagged statement, decide whether it's presented as fact or as an interpretation, and mark it for revision if it's written declaratively with no framing.
Rewrite each flagged statement using framing language that matches the strength of the evidence, such as “in this reviewer's opinion” or “the available documentation suggests.”
Where a record is silent on a point, make sure the conclusion acknowledges the gap rather than treating the silence as confirmation.
Read all the revised conclusions together within each section and confirm the framing is consistent, so declarative and opinion-framed statements don't sit side by side without a clear signal.
Problem: One section of the report states something that another section — written by the AI — states the opposite of.
Likely cause: Conclusions were added without cross-referencing them against the AI-generated content already in the report, or the expert's assessment simply differs from the AI's interpretation.
Example: An expert's conclusion states a fracture healed without complication, while an AI-generated section elsewhere in the report describes an ongoing complication from the same injury.
Fix:
Before writing any conclusions, read through the complete AI-generated report from start to finish.
As you write each conclusion, cross-reference it against the relevant sections of the AI-generated report to confirm it's consistent, or explicitly addresses a difference.
If a conclusion differs from the AI-generated content, use Files, AI Chat, or Search to determine which interpretation the records actually support before adding anything.
If the AI-generated content is incorrect, correct it using the source documents before adding your conclusion.
Once all conclusions are added, read through the full report one final time to confirm no section contradicts another.
Problem: A conclusion misses context that appears later in the report, such as a contradicting finding or an additional diagnosis.
Likely cause: Conclusions were written after only a partial reading of the report, before every workflow had finished and the full document had been read.
Example: An expert writes a prognosis conclusion after reading the first half of the report, missing a specialist opinion introduced later that changes the interpretation of the case.
Fix:
Before adding any conclusions, confirm that all workflows have finished running.
Open the report and read through it completely from start to finish without skipping sections, taking notes on key findings and any ambiguous or incomplete records.
Use Search or AI Chat to verify any unclear finding against the source documents before drawing a conclusion from it.
Write your conclusions based on the complete picture of the case, not on individual sections read in isolation.
After adding your conclusions, read through the report once more to confirm nothing introduced later in the document contradicts what you've written.
Problem: A reader without a clinical background can't understand what a conclusion means or why it matters to the case.
Likely cause: The conclusion was written in dense clinical language without an explanation of what the finding means in context.
Example: A conclusion references a specific diagnostic term with no explanation, leaving an attorney unable to tell how significant the finding actually is.
Fix:
After writing each conclusion, read it from the perspective of a non-medical reader, such as an attorney or paralegal without clinical training.
Identify any clinical term, abbreviation, or diagnostic concept the reader might not follow.
Rewrite the conclusion in plain language where possible.
Confirm that the significance of the conclusion to the legal argument is explicit.
Where simplifying a term would lose a clinically important distinction, keep the term and add a brief explanation instead of dropping the nuance.
Tip Read the entire report before writing a single conclusion, and read it again after. Most contradictions and missed-context problems trace back to a conclusion written from a partial view of the case. |
Summary
Every conclusion you add is now clearly separated from the AI-generated content, framed as professional opinion rather than fact, checked against the rest of the report for consistency, written only after a full read-through, and understandable to a reader without clinical training.