01 // the pipeline

From case brief to cross-examined report

case analysis

Your brief and domain are ingested. The engine performs case analysis, matching precedents and technical standards.

report hardening

AI analyzes the expert report structure against court standards. Claims are checked for citation backing. Opinions are evaluated for methodology that may face Daubert scrutiny.

cross-examination simulation

Opposing counsel simulation tests vulnerabilities across six axes: qualifications, methodology, basis, citations, logic, and bias indicators.

$ axion-legal --case-brief patent_brief.md --domain ip

[brief]     loaded: patent_llm_finetuning_brief.md
[domain]    ip — intellectual property
[analyze]   expert report... done
[verify]    DOIs: 2/15 valid (13 flagged — arXiv + fabricated)
[cross-exam] opposing counsel simulation... done

output: ip_expert_report_CROSS_EXAM.md
survivability score: 4/10

02 // illustrative output

Illustrative case study: patent infringement

Illustrative example. TechCorp v. NeuralScale is a synthetic patent-dispute scenario used to demonstrate typical Axion-Legal output patterns. Cross-examination found vulnerabilities that could significantly weaken the report in court.

case details

Illustrative case: TechCorp Inc. v. NeuralScale AI Ltd.

Subject: LLM Fine-Tuning Patent ('157 Patent)

Domain: Intellectual Property

Report size: 172 KB

pipeline results

DOIs verified: 9/9 valid (100%)

Survivability: 3/10 — FUNDAMENTALLY FLAWED

VULNERABILITY SCALE: 1 = MINIMAL RISK → 10 = CRITICAL EXPOSURE

Scores are illustrative. See the verification methodology for the scoring approach.

Qualifications

6/10 — vulnerable

Methodology (Daubert)

8/10 — critical

Basis / Findings of Fact

7/10 — high risk

Citation Integrity

7/10 — high risk

Logic and Reasoning

9/10 — critical

Bias Indicators

5/10 — moderate

Overall Survivability

3/10 — fundamentally flawed

1 minimal 5 moderate 10 critical

Key vulnerabilities found:

critical flaws (3)

Applied wrong legal standard (pre-AIA instead of post-AIA) and cited repealed statute §102(g) — disqualifying error

All infringement opinions based on reviewing only 15% of source code with no independent testing

Damages and Claim 12 opinions lack adequate evidentiary basis — framework without data

structural weaknesses (3)

False Daubert compliance — claims testability but never tested anything, wrong error rate cited

Unjustified confidence (90%+ certainty from 15% evidence) — could weaken under cross-exam

Expert exceeded scope — rendered legal conclusions on claim construction and patent validity

Disqualifying legal error found. Opinions built on 15% of evidence. Confidence levels that could weaken under deposition. The illustrative 3/10 survivability score means: revise before filing.

03 // economics

The cost of finding vulnerabilities early

<5 min initial signal in this run

4/10 survivability caught

6 attack vectors tested

~12min full pipeline time

Context: Expert witness reports and mock cross-examinations can require substantial professional time. A Daubert challenge that excludes your expert can significantly weaken your case. Axion-Legal does not replace your expert or legal team — it stress-tests the report before opposing counsel does.

04 // domains

Available expert domains

Attack strategy adapts to domain-specific standards and common vulnerabilities.

medical clinical standards, FDA, causation

engineering product liability, failure analysis

financial damages, valuation, lost profits

ip patents, trade secrets, Daubert

environmental contamination, exposure, EPA

general any domain, standard methodology

what you gain

Reasoning gaps caught before deposition.

what it costs

Not legal advice. Research support only. Requires bounded material and expert review.

Expert report, cross-examined before the courtroom.