Tutorial ID | T-06 |
Section | Preparing a Case/Reviewing the case |
Title | Starting Work on a New Case |
Subtitle | Confirm your documents, understand the claim, and set up the case before working on it. |
Before You Begin
You have created the case with setting proper Event date, and Case Details.
You know, at a high level, what the case is about.
All necessary documents have finished uploading, with none still in progress.
What You Will Accomplish
By the end of this tutorial, every new case you start will begin from a well-understood foundation. You will know how to confirm your documents, get an overview of the claim and set up the right access before you dive into detailed review.
Why Does this Tutorial Matter
A case built on a complete, clean, well-organized set of documents makes every later step — finding key pages, generating report sections, validating them — faster and more reliable.
What to Do if Something Goes Wrong
The list below covers the most common problems with starting a new case. Each entry follows the same pattern: what you will notice, why it likely happened, and how to fix it.
Problem: You begin building a chronology before the document set is complete.
Likely cause: Records were still arriving, and work started before the last of the document had been uploaded.
Example: You start working on a chronology report, even though several records related to the date of the accident haven't arrived yet. Records tied to the event date are exactly the ones a chronology depends on most, so starting without them risks building the case's central timeline on an incomplete picture.
Fix:
Click the case portal icon and check the case's status. If records are still expected, the case should be in Waiting for Records rather than Open.
Open Files and confirm which records are expected and which have actually arrived.
Review the document names and note the gaps — missing providers, missing date ranges.
Once you've confirmed the document set is complete, start working on the case.
Problem: You begin running workflows before the document set is complete.
Likely cause: When you are eager to get a quick read on the case, it is easy to run a workflow without stopping to check whether every expected document has actually arrived.
Example: A workflow generates a treatment summary for a case that's still missing two providers' records. The summary reads as complete and confident, but it has to be rerun entirely once those records arrive, since the missing treatment changes the picture.
Fix:
Confirm the case's status. If records are still outstanding, it should sit in Waiting for Records, not Open.
Open Files and check that every expected record has actually arrived before running a workflow.
Review the document names and note the gaps — missing providers, missing date ranges.
Once you've confirmed the document set is complete, start running workflows.
Problem: The main set of tasks chosen for the case is not the most efficient approach, given its size.
Likely cause: A default approach gets used without weighing it against the case's actual page count.
Example: A 40-page case is worked by generating a full AI report first and then validating it, when manually adding the small number of pages to the Workspace with the correct dates would have been faster and just as accurate. Conversely, a 2,000-page case is worked page by page, when generating an AI report first and validating it would have surfaced the case's main points far more quickly.
Fix:
Before working through case documents, evaluate the total number of pages in the case.
If the page count is small, start by adding pages directly to the Workspace, setting the correct date and other details as you go — this makes the case easy to navigate and validate from the start.
If the page count is large, start by generating an AI report to quickly surface the main points of the case, then validate that output against the source documents.
Choose your approach based on case size before starting substantive work, rather than defaulting to the same method every time.
Problem: The prompts used to run a workflow are not suited to this type of case.
Likely cause: A workflow is run using a default prompt, without checking whether its content actually fits the case's specific claim type, issues, or documentation.
Example: A workflow for a car accident case is run using a prompt written for a burn injury claim. The resulting output doesn't address the issues the case actually needs covered.
Fix:
Before running any workflow, get clear on the aims of the case, so you know what the output actually needs to accomplish.
Click AI Prompt Hub in the toolbar on the right of the screen.
Review the list of available prompts and identify the ones relevant to your case.
Check each prompt's content carefully to confirm it fits the case's claim type and issues before running it.
If none of the available prompts fit your case, click Create Prompt to build one suited to your case's needs.
Problem: The right people don't have access to the case from the start.
Likely cause: If the right people don't have appropriate access from the start — work is delayed, or exposed. Only Admin and Staff users can create and edit cases or run AI features; Presenters get a read-only view limited to cases they're explicitly assigned to.
Example: A presenter is given full Staff-level access to a case containing sensitive settlement details.
Fix:
Decide who needs to work on the case and at what role — Admin, Staff, or Presenter.
If a team member doesn't have an account yet, invite them from User Management and assign the correct role.
Use Share access, in the case's three-dot menu, to give specific users or presenters access to this case.
Confirm that sensitive material is only visible to those who should see it.
Summary
Your case now starts from a complete, clean, well-understood foundation. The documents are confirmed and organized, the issues and baseline are identified, and the right people have the right level of access.