Skip to content
Extract key testimonyDiscovery responses

Discovery Gap Finder

Flag every evasive, incomplete, or boilerplate-objection answer against what was actually asked — and walk away with your meet-and-confer list drafted.

Report a correction

Worked example · Synthetic fictional details

Sample prompt

Synthetic example: The names, dates, case details, and issues below are fictional CasePrompts sample values, not client data or evidence from a real matter. They demonstrate the prompt format, not the accuracy of any AI model's output.

Preview the complete sample prompt
Customize in builder
CONTEXT:
You are a litigation support analyst assisting plaintiff's counsel in a construction defect matter. The attached materials are the opposing party's discovery responses, paired with the requests and interrogatories they answer. The issues that matter are: subcontractor scopes of work; water intrusion repairs; notice to the GC.

INSTRUCTIONS:
Review every request/answer pair in sequence. For each, compare what was asked against what was actually answered, and classify any deficiency as exactly one of:
1. EVASIVE — the answer responds to a different question than the one asked.
2. INCOMPLETE — the answer addresses part of the request and is silent on the rest; identify the unanswered part.
3. BOILERPLATE OBJECTION — objections asserted with no substantive response, or "subject to and without waiving" followed by a non-answer.
4. IMPROPER REFERENCE — answers that point to documents or other responses without identifying them with specificity.
Answers that respond fully belong in a separate list, not the deficiency table. Do not argue whether objections are legally valid — identify what was asked and what was not answered.

TEMPLATE — format your output exactly as follows:
## Deficiency map
| Req # | What was asked (one line) | What was answered (verbatim key language) | Deficiency type | What a complete answer would cover |
|---|---|---|---|---|
## Meet-and-confer points
Numbered list, one per deficient response, phrased neutrally and ready to paste into a letter, each with the request number.
## Answered in full
Request numbers only.

EVIDENCE REQUIREMENTS:
For every finding, cite the request/interrogatory number and the page or Bates number of the response. Quote the operative answer language verbatim — a response may not be characterized as evasive or incomplete unless its own words support it. If a referenced document or verification page is missing from the production, list it under a final heading "Missing from the production" rather than assuming its contents.

If the provided record does not support a requested finding, write "not found in the record." Do not infer or fill gaps from outside knowledge.
The scope is limited to the requested task, stated issues, and provided record. Do not introduce outside subjects or sources.
A human reviewer must verify every material statement and citation against the source record before relying on or sharing the output.

AI output is a starting point, not work product. Verify every citation against the record before you rely on it, file it, or send it.

What you'll fill in

  • 01Your role
  • 02Case typee.g., "trucking liability," "first-party property," "medical malpractice"
  • 03Key issue(s) — 1 to 3e.g., "vehicle speed at impact; brake maintenance; visibility"

Pro tip Run it the day responses land — the meet-and-confer clock starts whether or not anyone has read the responses yet.

Record-first path

Related verification workflow