#Connect and identify the case
Connect ElevenFlo MCP with the setup guide. In Claude or
ChatGPT, begin with the debtor name and any known court, case number, or
petition date. Ask your client to call find_cases and confirm the returned
case_id before it analyzes filings. This prevents a similarly named debtor
or an earlier case from entering the research record.
ElevenFlo has 2,000+ fully covered cases. Full coverage means the case includes searchable filing text in addition to docket metadata. It does not promise that every exhibit is machine-readable or that every restricted document is available.
#Bound the docket review
State the question and the date or docket range. For example, ask for filings
since a stated hearing, entries related to a DIP milestone, or activity between
two known dates. Then ask your client to call list_docket_entries for a
recent bounded list or search_filings for a focused issue.
Keep metadata and document evidence separate. A docket entry can establish
that a filing appears on the docket. It may not supply the operative language
inside the filing. Use list_document_types when the filing category is
unclear, and retain the returned document_id values for the next step.
#Read the supporting text
Use search_document_summaries or get_document_summary to orient the
review, not to settle an exact term. For the language that supports the answer,
call search_document_chunks across the selected documents. Follow with
read_document_chunks for chosen chunks or extract_document_passages
for a focused question within one document.
Ask the client to cite the docket number, filing title, filing date, and source link for every material point. It should distinguish a motion from a proposed order, an interim order from a final order, and requested relief from relief the court granted. If the requested document or text is unavailable, the answer should say so instead of filling the gap from model memory.