Lexis, Westlaw, vLex. Your legal research history is being sold.
Short answer
Every search you run on Lexis, Westlaw, vLex, or Bloomberg Law is logged, retained, and in some configurations sold downstream. Your research history reveals what you are working on, who your client is by inference, and which arguments you are testing. The professionals most exposed to confidentiality obligations are also those leaving the deepest research trail of any white-collar field.
What the research platforms log
The four major legal research platforms (LexisNexis, Westlaw owned by Thomson Reuters, vLex which absorbed Fastcase in 2023, and Bloomberg Law) all log search activity at the user account level. The logs are mandatory infrastructure: they underpin metered billing, the analytics products platforms market back to firms, and the AI features now layered on top.
The captured data, across all four platforms, includes: every search query you typed, every case you opened, every duration spent on each document, every annotation you made, every download, every print, every export. Account metadata ties the activity to your name, your firm, and (usually) your bar number. Time stamps are precise to the second.
The retention period varies by contract. Default retention on standard firm contracts is between two and seven years. Retention on enterprise contracts with explicit privacy addenda can be shorter. Retention on individual subscriptions is generally longer. Solos and small firms have the longest retention exposure of any segment.
Where the data goes
Three flows. Each one matters for a different reason.
The platform’s own analytics products
The platforms sell data products built on aggregated user behavior. Westlaw’s Litigation Analytics, Bloomberg Law’s Litigation Analytics, and Lexis’s Context tool are marketed to firms as litigation intelligence. The products use aggregated data: which cases get cited together, which arguments succeed before which judges, which firms research which areas. The aggregation typically de-identifies individual users. The aggregation is also reverse-engineerable in narrow circumstances.
LexisNexis Risk Solutions
LexisNexis Risk Solutions is the same parent company as the legal research platform but operates a different product line: data brokerage, identity verification, and bulk records sold to investigators, insurance companies, and government agencies. The corporate structure separates the legal research data from the risk-solutions data. The structural separation does not always survive subpoenas. Several documented cases have shown research-platform metadata reaching the broker side under specific legal demands. The framework around third-party data flow is the same one we cover in Plaid and the financial data brokers: corporate boundaries are weaker than user expectation.
Subpoena and discovery exposure
Your research history can be subpoenaed and has been in documented cases. Opposing counsel in a high-stakes commercial matter requested the lead attorney’s Westlaw search history during the relevant period. The court ruled the request overbroad but allowed a narrowed version, which produced a list of search queries that revealed the firm’s strategic thinking. The exposure is rare in absolute terms but devastating when it lands, because the queries themselves are work product.
What the AI features changed
Each platform shipped an AI assistant in 2024 or 2025 (Lexis+ AI, Westlaw Precision with CoCounsel, Bloomberg Law’s AI features, vLex Vincent AI). The assistants take natural-language queries, search across the platform’s case database, and synthesize answers with citations.
The AI conversations are an additional layer of logging. Your prompt to the AI is more revealing than your raw search queries because it is articulated in plain language. “Has any state court accepted the argument that an employer’s monitoring software constitutes a wiretap when the employee did not sign a specific consent” is a query that reveals the case theory directly. The same query as a Boolean search across cases is less interpretable to anyone reading the log.
The AI feature retention is generally the same as the underlying platform retention, plus contractual terms with the AI vendor (typically OpenAI or Anthropic on the back end). The AI vendor terms are the same posture we walked through in your law firm uses ChatGPT. The legal research platform AI is contractually closer to the enterprise tier of the underlying chatbot, but the underlying privilege analysis is the same.
What firms have done about it
1. Negotiate retention in the contract
The default retention period is the platform’s preference, not yours. Larger firms now negotiate retention of less than ninety days for active matters and immediate purge after matter closure. The platforms accept these terms when the firm is large enough to push. Small firms accept the default. The conversation is worth having even when you think the answer will be no.
2. Use account compartmentation per matter
Westlaw and Lexis allow firms to maintain matter-specific accounts. Research for one matter goes through one account. Research for another matter goes through a different account. The cross-matter linking is broken at the platform level. The trade-off is administrative overhead. For high-stakes matters, the overhead is worth it.
3. Treat AI prompts as work product, audit-ready
Train associates to formulate AI prompts that read as work product, not as case admissions. A prompt phrased as “the strongest counter-argument to our position is” yields a useful answer and a damaging log entry; rephrased as “research the published responses to argument X” it yields the same answer with a far less revealing trail.
4. Audit the firm’s account access
Periodically review who at the firm has access to the research platforms. Departed attorneys, former contractors, and shared accounts that have not been deprovisioned create exposure that the firm has not measured. The deprovisioning hygiene is the same one we covered for general account management in operational identity separation between research accounts and matter-level activity.
Frequently asked questions
Does using the platform from a private browser session help?
No. The logging is server-side, tied to your account, not to your browser. Incognito mode and VPN only affect the local browser state and the network path, neither of which touches what the platform records on its side. Authentication to your account is the trigger for logging, regardless of how you connected.
Can I export my own research history?
Yes, on most platforms. Lexis and Westlaw have a research-history view that lets you export the last several months of activity. The export is useful for billing, for handing off matters, and for understanding what is in the platform’s records. Export quarterly. Read what you find. The exercise itself is illuminating.
Can I delete my research history?
Sometimes, partly. The user-facing deletion option (where it exists) removes items from your view. The platform’s internal logs are governed by the contract. CCPA and GDPR rights apply to user data on these platforms but the deletion right is partially carved out for records the platform claims it must retain for service-integrity or legal purposes. Practical deletion is contractual, not regulatory.
What about open access tools like Google Scholar and CourtListener?
Different exposure profile. Google Scholar logs queries against your Google account if signed in. CourtListener (run by Free Law Project, a non-profit) has narrower retention and a non-commercial mission. For preliminary research and for queries you do not want logged in a commercial database, CourtListener is the cleaner option. Coverage is narrower than Westlaw or Lexis, sufficient for many tasks.
There’s no perfect setup. Anyone selling you perfect is selling fear. The goal is simple: make yourself a harder target than the person next to you.
