From Answer to Evidence: How a Legal Professional Uses Scoopie’s Citations
For one legal professional who reviews witness testimony, a useful AI answer needs to lead somewhere: back to the record.
During our interview, they described the feature in straightforward terms:
“I click the note, and it shows me where in the text file the support for the claim is.”
That connection mattered because their work did not end with understanding what a witness had said. If they relied on a factual statement in a filing, they needed to support it with a citation to the record.
Scoopie helped them move from a question to an answer, then from that answer to the supporting passage. The final review remained theirs.
This article is based on a real customer interview. Names, affiliations, locations, specific job titles, and identifying case details have been omitted. It describes an individual’s workflow, not an organization-wide deployment.
A summary was a starting point, not something to cite in a filing
The user worked with extensive witness testimony. They needed to understand each person’s involvement in a case, identify relevant statements, and locate the records supporting those statements.
They prepared text versions of the testimony and connected those files to Scoopie. They could then ask about an individual witness or select several witnesses’ records to examine what they had said about an event.
The answers helped them get oriented. But a summary alone could not complete the task.
Knowing that a witness had described something was different from locating the testimony they could actually rely on. The user needed to return to the record and inspect the supporting material.
That is where the clickable citations became useful.
The citation gave them a place to look
The user described a direct connection between Scoopie’s answer and the text behind it. When they clicked a citation, they could see where the connected file supported the statement.
This made the answer a route into the material, rather than a separate piece of text to evaluate on its own.
The distinction is practical. An AI can provide an account of what a document says, but the person reviewing it still needs to examine the document. In this user’s workflow, the citation helped bridge those two steps.
They could begin with a question such as what a particular witness had said, read Scoopie’s response, and follow the source reference to inspect the relevant passage.
Their description of the feature was about finding support for a claim. It was not a claim that every AI answer was automatically correct.
It fit a requirement already built into their work
The user explained that factual claims in their filings needed citations to the record.
That requirement existed before Scoopie entered the workflow. Using AI did not remove it.
What Scoopie offered was help with the work leading up to that citation: finding relevant testimony, organizing what it said, and making the supporting text accessible from the answer.
The app’s source reference and the record citation used in a filing served related but different purposes. The first helped the user locate and inspect the material. The second had to be grounded in the record they ultimately relied on.
For this user, that made source access more than a reassuring symbol next to an AI-generated sentence. It supported a task they already had to perform.
They still checked the original PDF
The interviewee was explicit about the final step:
“Before I file anything, I’m going to manually verify it against the PDF.”
Their workflow involved two distinct kinds of source checking.
First, they followed Scoopie’s citations into the text files they had connected. Then, before submitting a filing, they checked the material against the original PDF.
That distinction matters. It would be inaccurate to describe this experience as Scoopie automatically verifying a filing or eliminating the need to consult the original record.
The user had prepared the text files themselves. They also described comparing those files with the PDFs during preparation to check for transcription problems. The AI-assisted review sat within that broader process.
Scoopie helped them find and understand material. It did not become the final authority on what the record contained.
The benefit was easier review, not less responsibility
The user described Scoopie’s analysis as making their work more efficient. Their account shows where that usefulness came from.
They could ask a focused question instead of relying only on a first reading to identify the relevant material. They could follow the answer back to supporting text. They could then carry out their own check against the original.
The interview does not establish a measured reduction in review time or a guarantee of accuracy. It does show how source access can make an AI answer more usable in a workflow that requires verification.
The user did not have to choose between accepting the summary and ignoring it. They could use it to guide the next stage of review.
A useful answer gives you somewhere to go next
For this legal professional, Scoopie’s value was not simply that it could explain what the records said. It was that the explanation came with a way to inspect the supporting material.
The answer helped them get oriented. The citation helped them find the passage. The original record remained the basis for their final check.
That is a useful way to evaluate an AI document tool: after it answers your question, how easily can you examine the source?
Scoopie lets users ask questions about connected files and open cited passages for review. Start with a public document or a file you are authorized to use. Ask a question, open the citation, and compare the answer with the source.
The goal is an answer you can work with because you can check it.
Sources
Scoopie customer interview transcript. Identifying details and the original source link are withheld to protect the interviewee’s anonymity. Quoted excerpts have been lightly edited for punctuation and readability.
Scoopie: Official Product Information, consulted for the description of connected-file questions and source viewing.