The Problem with Traditional Document Search
General AI tools work from what fits in one conversation, and you have to know which documents to give them.
But legal work requires cross-document analysis:
- "Did the plaintiff's testimony match the medical records?"
- "Which expert witnesses disagree about causation?"
- "What did the surgeon say in deposition versus what the operative report shows?"
- "Find all evidence of pre-existing back pain across 14 years of medical records"
Answering these questions manually requires:
- Reading multiple documents (depositions, medical records, expert reports)
- Remembering key facts from each document
- Comparing statements across documents
- Synthesizing findings into a cohesive answer
This takes hours or days for a human paralegal. General chat tools lose track across thousands of pages and dozens of files.
How Deep Agent Solves This
Deep Agent is AI that works like an experienced paralegal. When you ask a complex question, Deep Agent:
- Understands your question: "Find contradictions between Dr. Smith and Dr. Jones about causation"
- Plans a search strategy: "I need to find Dr. Smith's deposition, Dr. Jones' expert report, and any medical records where they discussed causation"
- Searches all relevant documents: Opens Dr. Smith's deposition (300 pages), Dr. Jones' expert report (200 pages), and medical records (12,000 pages) simultaneously
- Compares findings across documents: Identifies where Dr. Smith said one thing and Dr. Jones said something different
- Provides a synthesized answer with citations: "Dr. Smith testified the surgery caused permanent nerve damage (Deposition p.234), but Dr. Jones concluded pre-existing degenerative disease was the primary cause (Expert Report p.45)"
Runs in the Background
Hand the Deep Agent a multi-step question, keep working, and come back to a cited answer. Queue several questions at once; LexHound emails you when processing is done, and every answer shows its sources.
Example: Cross-Document Analysis in a Personal Injury Case
Example workflow, not a client matter.
Personal Injury Case: Auto Accident with Disputed Pre-Existing Condition
Documents:
- Plaintiff's deposition (450 pages)
- 14 years of medical records (23,000 pages)
- Defense medical expert report (180 pages)
- Plaintiff's treating physician notes (2,500 pages)
Attorney Question:
"Did the plaintiff ever report back pain before the accident date of March 15, 2024?"
Why This Question Is Complex:
- Requires searching 23,000 pages of medical records spanning 14 years
- Must compare dates (pre-accident vs post-accident)
- Must identify synonyms ("back pain," "lumbar pain," "lower back discomfort," "sciatica")
- Must cross-reference plaintiff's testimony with medical records
Deep Agent Answer:
Found 3 instances of pre-accident back pain complaints:
1. October 12, 2018 - Dr. Williams Office Visit
"Patient reports occasional lower back stiffness after yard work. No radiating pain. Resolved with OTC ibuprofen."
(Source: Medical Records, Page 12,456)
2. January 8, 2021 - Emergency Room Visit
"Chief complaint: Acute lumbar strain after lifting heavy box. X-ray shows no fracture. Discharged with muscle relaxers."
(Source: ER Records, Page 18,234)
3. June 3, 2023 - Annual Physical
"Patient denies any chronic pain. No back pain reported."
(Source: Medical Records, Page 21,890)
Cross-Reference with Deposition:
Plaintiff testified: "I never had back problems before the accident." (Deposition p.89)
CONTRADICTION: Deposition testimony conflicts with 2018 and 2021 medical records showing prior back pain complaints.
Manual Research Time Estimate:
- Read 23,000 pages of medical records: 60-80 hours
- Create timeline of back pain mentions: 8-12 hours
- Cross-reference with deposition: 4 hours
- Total: roughly 72-96 hours of reading and cross-referencing
General AI Chat Tools vs. LexHound Deep Agent
| Capability | General AI chat tools | Deep Agent (LexHound) |
|---|---|---|
| Which documents it works from | What you paste or upload into one conversation | Your whole case file, already indexed |
| Compare testimony across depositions | You pick the excerpts and compare them yourself | Compares across every deposition you uploaded |
| Timeline across years of medical records | Limited by how much fits in one conversation | Builds the timeline from the full record, on request |
| Cross-reference deposition with medical records | You read both and compare | Flags contradictions, citing both sides |
| Multi-step questions | You run each step yourself | Plans the steps and runs them in the background |
| Citations | Not reliably tied to page and line | Exact document, page, and line, with the passage highlighted |
Specific Use Cases Where Deep Agent Excels
Medical Malpractice
- Timeline Reconstruction: "Create a timeline of all pain complaints from admission to discharge"
- Standard of Care Analysis: "Compare the nursing notes with the hospital's fall-prevention protocol (uploaded) and list deviations"
- Expert Comparison: "Where do plaintiff's and defense's experts disagree about causation?"
- Documentation Gaps: "Find all instances where nursing notes mention events not in physician notes"
Personal Injury
- Pre-Existing Conditions: "Find all mentions of back pain before the accident date"
- Witness Impeachment: "Compare defendant's police report statement with deposition testimony"
- Damages Verification: "Total the medical expenses across all billing records" (verify totals against the cited bills)
- Treatment Gap Analysis: "When did plaintiff stop treating for this injury?"
Product Liability
- Multi-Plaintiff Patterns: "Which plaintiffs experienced the same side effect?"
- Temporal Analysis: "When did each plaintiff start using the product relative to symptom onset?"
- Manufacturer Knowledge: "What did the manufacturer know about this defect before recall?"
How Deep Agent Provides Citations
Every answer from Deep Agent includes exact citations:
- Document name: "Dr. Smith Deposition"
- Page number: "Page 234"
- Line numbers (for transcripts): "Lines 15-18"
- Exact quote: Full text of relevant passage
This means you can verify every fact before using it in court. Deep Agent shows its work, so you can check every fact against the source before you rely on it.
Frequently Asked Questions
What does "Deep Agent" mean in simple terms?
"Deep Agent" means AI that can think through complex questions requiring multiple steps. Instead of just finding keywords, it plans a research strategy, searches multiple documents, compares findings, and synthesizes an answer with citations, and it runs in the background while you keep working.
How many documents can Deep Agent analyze at once?
Unlimited. LexHound has handled case files of more than 200,000 pages, and the Deep Agent searches across the entire set. The more documents you upload, the more comprehensive the analysis.
Does Deep Agent work with handwritten notes or poor-quality scans?
Deep Agent works on text extracted from documents. LexHound reads scanned and faxed records with OCR (Optical Character Recognition). Typed text reads reliably; handwriting is harder, so spot-check messy handwritten notes against the cited page.
Can I trust Deep Agent's citations?
Always verify. Every answer includes the exact document, page, and line (for transcripts), with the passage highlighted, so you can check it in one click before using it in court. Think of Deep Agent as a research assistant: it finds leads, you verify them.
How is Deep Agent different from ChatGPT's file upload feature?
General chat tools can summarize files you upload, but they aren't built to track every person, date, and claim across a whole case file or to cite each fact to its page. Deep Agent is purpose-built for cross-document legal analysis, with citations you can check.
Does Deep Agent understand legal citations and case law?
Deep Agent is optimized for fact-finding in documents (depositions, medical records, expert reports), not legal research (case law, statutes). For legal research, use tools like CoCounsel (formerly Casetext) or Westlaw. For document analysis and fact investigation, use LexHound's Deep Agent.
Can Deep Agent draft legal documents or briefs?
Deep Agent is designed for research and analysis, not document drafting. It can find facts and citations to support your arguments, but you write the brief. Some attorneys use Deep Agent to gather evidence, then copy the findings into Word or their drafting tool.
How long does Deep Agent take to analyze a large case file?
Processing runs in the background, and we email you when it's done. Simple questions come back quickly, and multi-step requests run in the background while you work.
Can I export Deep Agent's findings for use in trial?
Yes. All findings can be exported to PDF or Word (DOCX) with citations intact. Many attorneys use Deep Agent's output as the foundation for deposition prep, trial notebooks, or expert witness cross-examination outlines.
Is Deep Agent available on all LexHound plans?
Yes. Deep Agent is included with every plan and every seat at $999/seat/month. You get the full power of cross-document analysis regardless of subscription tier.