1. The Scenario
A software engineer at a mid-size SaaS company wants to leave and join a competitor. She signed a non-compete agreement when she was hired two years ago. The agreement restricts her from working for any "competing business" within a 50-mile radius for 18 months after departure.
The partner asks you to research whether this non-compete is enforceable under California law (where the employee lives and works) and, alternatively, under Texas law (the governing law specified in the agreement). She needs a memo by end of day.
Key details:
- Employee is a software engineer (not C-suite or executive)
- Non-compete: 18 months, 50-mile radius, "competing business"
- Employee works in California; contract says Texas law governs
- No sale of business involved — employment context only
- Company has no trade secrets claim at this stage
2. The Traditional Approach
A traditional legal research memo on non-compete enforceability involves multiple research tracks:
Statutory research — California (1-2 hours)
Research California Business & Professions Code Section 16600, the 2024 amendments (AB 1076, SB 699), and their combined effect on non-compete enforceability. Pull relevant legislative history.
Case law research — both jurisdictions (2-3 hours)
Research California case law (Edwards v. Arthur Andersen, etc.) and Texas case law (reasonableness requirements). Shepardize key cases. Research choice-of-law disputes between California and Texas.
Analysis and memo drafting (2-3 hours)
Synthesize research into a structured memo with IRAC-style analysis for each jurisdiction. Draft conclusion and practical recommendations.
Review and cite-check (1 hour)
Verify all citations, check that cases are still good law, proofread the analysis, and ensure the memo answers the partner's questions.
Traditional Time Estimate
6-9 hours of focused attorney time for a thorough two-jurisdiction research memo.
3. The AI-Assisted Approach
AI changes this workflow in two critical ways: it accelerates the initial research phase and produces a structured first draft. But verification becomes even more important because AI can fabricate legal citations.
Run the Research Prompt (5 min)
Use the prompt below with a Deep Research tool (Gemini, ChatGPT, Perplexity) or a standard AI with web search. The key: ask it to provide specific citations with case names, dates, and reporters so you can verify them.
Verify Every Citation on Westlaw/Lexis (45-60 min)
This step is non-negotiable. For each case the AI cites, confirm: (a) the case exists, (b) the holding is accurately stated, (c) it is still good law. Run Shepard's or KeyCite. This is where AI research fails most — fabricated or mischaracterized citations. Budget real time here.
Targeted Follow-Up Research (20 min)
Use the AI to explore specific angles: "What is the current status of FTC enforcement regarding non-competes?" or "Are there any 2024-2025 Texas cases that narrowed enforceability?" Cross-reference with your Westlaw/Lexis results.
Draft the Memo Using AI as a Framework (30-45 min)
Use the AI's output as a structural framework for your memo. Rewrite the analysis in your own words, replace any incorrect citations with verified ones, add your own legal judgment, and ensure the conclusion addresses the partner's specific questions.
Final Quality Check (20 min)
Read the complete memo as if you had written it from scratch. Does the analysis flow logically? Are the citations properly formatted? Does the conclusion follow from the analysis? Would you be comfortable presenting this to the partner?
4. The Prompt
This prompt is best used with a Deep Research tool (Gemini Advanced, ChatGPT Pro, or Perplexity) for the citation verification advantage. It can also be used with standard Claude or ChatGPT — but expect to verify more aggressively.
Complete Prompt — Non-Compete Research Memo
You are a senior employment litigation attorney preparing a legal research memorandum. Research and analyze the following issue thoroughly.
RESEARCH QUESTION:
Is the following non-compete clause enforceable?
- Employee: Software engineer (non-executive) at a mid-size SaaS company
- Restriction: Cannot work for any "competing business" within a 50-mile radius for 18 months post-employment
- Employee resides and works in California
- The employment agreement specifies Texas as the governing law
- No sale-of-business context; this is a pure employment non-compete
- No pending trade secret claims
ANALYSIS REQUIRED:
## 1. CALIFORNIA LAW ANALYSIS
- Analyze California Business & Professions Code Section 16600
- Address the 2024 legislative changes (AB 1076 and SB 699) and their effect on non-competes
- Cite the leading California Supreme Court and appellate decisions on non-compete enforcement
- Address whether California would enforce a Texas choice-of-law provision in an employment non-compete
- Analyze under the framework of Application Group, Inc. v. Hunter Group, Inc. and any subsequent developments
## 2. TEXAS LAW ANALYSIS
- Analyze the Texas Business & Commerce Code Chapter 15 (Covenants Not to Compete)
- Identify the reasonableness requirements under Texas law
- Assess whether 18 months / 50 miles / "competing business" satisfies Texas requirements
- Cite leading Texas Supreme Court cases on non-compete enforcement for non-executive employees
- Address whether a Texas court would reform an overbroad non-compete versus invalidating it
## 3. CHOICE-OF-LAW ANALYSIS
- Analyze which jurisdiction's law would likely apply
- Discuss California's strong public policy against non-competes and its effect on choice-of-law analysis
- Address SB 699's extraterritorial reach provisions
- Cite relevant federal and state cases on choice-of-law in employment non-compete disputes
## 4. CONCLUSION AND PRACTICAL RECOMMENDATIONS
- Provide a clear assessment of enforceability under each jurisdiction
- Recommend a litigation strategy for the employee
- Identify any additional facts that would strengthen or weaken the analysis
- Flag any pending legislation or regulatory developments (including FTC activity)
FORMAT:
- Use IRAC structure for each issue
- Provide full case citations (case name, reporter, year, court)
- Distinguish holdings from dicta
- Note the current procedural status of any cited cases
- Flag any areas where the law is unsettled or evolving
IMPORTANT: Provide specific, verifiable citations. I will independently verify every case and statute you reference.Why This Prompt Works
Specific jurisdiction framing
By specifying California and Texas explicitly and providing the exact choice-of-law conflict, the AI can focus research rather than surveying all 50 states.
Concrete factual details
The prompt includes role (software engineer), terms (18 months, 50 miles), and context (no trade secrets, no sale of business). This allows the AI to apply the law to facts rather than giving abstract summaries.
IRAC structure requirement
Requesting IRAC format produces output that closely mirrors a legal memo's structure, saving significant reformatting time.
Verification signal
The final line — "I will independently verify every case" — signals to the AI that accuracy matters. Research suggests this type of accountability framing improves citation accuracy.
5. The Review — Evaluating AI Output
Legal research is where AI hallucination carries the most professional risk. Here is a systematic review process:
Citation Verification Is Non-Negotiable
Check every single case citation on Westlaw or LexisNexis. Confirm: (1) the case exists, (2) the holding is accurately described, (3) it has not been overruled or distinguished on the relevant point. AI tools routinely fabricate case names, invent citations, and mischaracterize holdings. The consequences of submitting fabricated citations to a court are severe.
Verify Statutory Currency
Confirm that the statutes cited are the current versions. Non-compete law is evolving rapidly — California's 2024 amendments, the FTC's rulemaking, and state-level legislative activity mean that statutory analysis from even 12 months ago may be outdated.
Check the Analysis, Not Just the Citations
Even when the AI cites real cases, it may draw incorrect conclusions from them. Read the actual case to confirm the AI's characterization of the holding, and verify that the case supports the proposition for which it is cited.
Fill the Gaps
Run your own targeted searches on Westlaw or Lexis for issues the AI did not address. Did it miss any recent decisions? Did it overlook relevant secondary authorities (Restatement provisions, treatises, bar journal articles)?
6. The Result — Time & Quality Comparison
Traditional Approach
6-9 hrs
Full two-jurisdiction research, analysis, and memo drafting
AI-Assisted Approach
2-3.5 hrs
AI-assisted research and drafting with rigorous human verification
The Important Nuance
The time savings are real but come with a critical caveat: the verification phase cannot be shortened. The 45-60 minutes for citation verification is a minimum, not a target to optimize away. If anything, less experienced researchers should budget more time for verification because they may have less intuition about which citations "feel wrong."
Where the real efficiency gain lies is in the research discovery phase — finding relevant authorities, identifying the legal framework, and structuring the analysis. The AI handles these tasks in minutes rather than hours, freeing your time for the higher-value work of verification, judgment, and synthesis.
Try Another Application
Legal research is foundational, but AI assists across the full practice workflow. Explore demand letter drafting, due diligence, and more.
Ready for structured learning? Explore the Learning Program →
Comments
Loading comments...