Skip to content

Use caseFor tax and legal

Check the facts in your memo before you sign it.

Lenz is a fact-checking service for tax and legal memos. It identifies the factual statements it can check against independent public sources — what a statute requires, what a court held, a rate, a deadline, a form number — and returns a result for each. It also looks up the cases, statutes and EU acts the memo cites and checks each against its source. It is not a citator.

Private by default. Zero retention on paid Pro and Scale.

Example: a tax memo“A calendar-year C corporation files its federal income tax return on Form 1120, due April 15, and can extend the deadline by six months by filing Form 7004. An S corporation files the same Form 1120 by the same date.”

US federal, as of 2 September 2026

What Lenz found

FalseAn S corporation generally files Form 1120-S for its federal return.

Mostly FalseCalendar-year S corporations generally file federal returns by March 15.

01Example

Five statements from a draft memo, each one checked.

Deep check · Checked 2 September 2026.

Open a row for the sources it used and the reasoning that weighed them.

02Where you work

Check the memo where you write it.

Choose where you work 3 ways to run a check

Check the memo in Claude.

Connect Lenz to Claude with one sign-in and no key, and ask for a check before the memo leaves the window: “Check the claims in this memo with Lenz.” Hand over the memo whole; each statement comes back with its own result.

  1. Extract the statementsThe memo goes in whole; each statement comes back with its own result.
  2. Assess them in bulkA quick check returns a verdict and a confidence level for each statement, for triaging a long memo.
  3. Send the uncertain ones to a deep checkThe deep check returns the verdict with its sources and the reasoning.
  4. Route what needs a reviewer to oneA person reads what is left before the memo goes out.

Connect Claude

client-memo.docx

Check the claims in this memo with Lenz.

Lenz · quick check · 5 statements

Lenz checked five statements. Two need attention:

  • FalseAn S corporation files on Form 1120-S, not 1120.
  • Mostly FalseCalendar-year S corporations file by March 15.
  • True ×3Form 1120, April 15, Form 7004 extension.

Lenz · deep check · 2 statements · sources + reasoning

Illustration · Claude with the Lenz connector. Also in ChatGPT.

03What is checked

Lenz checks each statement against the public record.

A memo states things about the law before it argues from them: what a statute requires, what a court held, which rule governs and from when.

Lenz checks each against the published rule and the public record — what the memo asserts, not whether a rule applies to your client.

  1. Statutes and the rules made under them

    “The statute requires X.” “The rule took effect in March.” Lenz looks for the published text and the authority’s own guidance, and returns what it finds, including when it finds too little.

  2. Case law: verify what a court held

    A model can name a real case and misstate its holding, or cite one that does not exist. Lenz looks the citation up, reads the opinion and checks it against what the memo says the court held. When no public record settles a statement, Lenz answers “No Public Record” instead of a score. That doesn’t make the statement false.

  3. Forms, deadlines, rates and thresholds

    The mechanics the example above checks. They turn on the jurisdiction and the tax period, so a memo that names both gives Lenz something to check against; one that leaves them implied does not.

04Before you rely on it

AI tax advice and legal research, before you rely on it.

An assistant’s answer reads the same whether it is right or a year out of date, and the memo drafted from it carries the error forward.

Hallucinated holdings and legal AI accuracy

Lenz looks the citation up, reads the opinion, and answers whether the case held what the memo says it held. One it cannot find comes back “Not checked” with the reason, never as fabricated.

Statutes and EU acts it cites

A memo leans on a section or an act for what it asserts. Lenz reads the published text of the section, or finds the act in the EU’s own register, and checks the sentence against it. Cite a directive at a number the register holds as a regulation and it comes back “Not checked”, with the type and the number to correct.

Regulatory claims in a legal document

What a rule requires, when it took effect, the threshold at which it bites, who it covers: statements with a published answer, each returned with the regulation or guidance it was checked against. Whether the rule reaches your client is analysis, outside the check.

05Scope

What it checks, and what it doesn’t.

Checks

The factual statements the memo makes, against public sources.

The cases, statutes and EU acts the memo cites, and any source it links: Lenz finds each one, reads it, and checks that it says what the sentence claims. One it cannot find comes back “Not checked” with the reason.

  • Statutes
  • Case holdings
  • EU acts
  • Forms
  • Deadlines
  • Rates and thresholds

Doesn’t check

It is not a citator: it won’t tell you whether a case is still good law.

It does not check citation format, authenticate a quotation, or assess the quality of the legal analysis, and it does not judge whether a model wrote the memo.

  • Good law
  • Citation format
  • Quotations
  • Legal analysis
  • AI detection

06Record

A record you can keep.

Keep it with the memo, so “how was this checked?” has an answer: the claim as Lenz read it, the verdict and its score, every source named and dated, and the case made for and against it before the verdict was reached.

1–2min

Deep check

The record comes from a deep check. Lenz researches independent public sources, has models argue both sides, and returns the verdict with those sources and the reasoning, in about a minute or two per statement.

~15sec

Quick check

A panel of models returns a verdict and a confidence level in about fifteen seconds, for triaging a long memo. Those verdicts are reviewers’ notes, not checked sources — where search returns nothing, a reviewer may answer from what it already knows. When the deep check disagrees, rely on the deep check.

Verification recordChecked 2 Sep 2026

An S corporation has an April 15 federal income tax return filing deadline.

3/10 Mostly false
Suggested rewrite

A calendar-year S corporation has a March 15 federal income tax return filing deadline.

Sources · 17, each dated
  • uscode.house.gov26 U.S.C. § 6072(b)
  • govinfo.govretrieved 2 Sep 2026
  • ecfr.govretrieved 2 Sep 2026
  • +14 more
Reasoning
ForThe 15th is the right day of the month.
AgainstFor a calendar-year S corporation, the month is March.

Illustration · record fields abridgedlenz.io/c/…

07Walkthrough

See a memo checked, start to finish.

Everything above, in under a minute: one employment law memo going through Lenz.

Walkthrough · employment law memo · 58 s

Warranty

Vendors disclaim accuracy. Lenz warrants the verdict.

A verdict can qualify when a full verification returns True or False with high panel confidence and enough independent sources behind it. Its certificate is a document you can hand to the client with the memo.

On a paid Pro or Scale plan, every full verification that qualifies carries a contractual warranty, at no extra charge.

How the warranty works

08Pricing

Try free, no card.

Free

$0

100 quick checks or 10 deep checks a month, in any combination.

Try free
Plus

$7.99/ month

500 quick checks or 50 deep checks a month, in any combination.

Or $6.59 a month, billed $79 a year

Choose Plus
Pro · warranty included

$99/ month

5,000 quick checks or 500 deep checks a month, in any combination. The warranty on qualifying verdicts.

Or $83 a month, billed $990 a year

Choose Pro
Scale · warranty included

$399/ month

20,000 quick checks or 2,000 deep checks a month, for production volumes. The warranty on qualifying verdicts.

Or $333 a month, billed $3,990 a year

Choose Scale

09Questions

Before you send a client memo.

Not answered here? Talk to us

Published statutes, regulations and the tax authority’s own guidance, plus independent commentary. A deep check includes the sources it used and the reasoning behind the result.

Only what the statement says. Include the jurisdiction and the tax period; Lenz searches public sources for the rule applicable to that statement and returns what it finds.

US case law, through Lenz’s own index of CourtListener citations; US Code sections and federal regulations, read from their published text; EU regulations, directives and decisions, identified through a register synced daily from the EU's own open data and read on EUR-Lex; UK Acts of Parliament cited by title and section (“section 172 of the Companies Act 2006”) and statutory instruments cited by number, read on legislation.gov.uk; and German federal law cited by paragraph or article (“§ 823 BGB”, “Art. 3 GG”), read on gesetze-im-internet.de. A citation it recognizes but cannot find comes back “Not checked” with the reason, never as wrong. For UK and German law, when there are signs that the memo means a different version of the law than the one on the page, such as a repeal or a note on changes, a mismatch comes back “Not checked” rather than as an error. State statutes are not read.

Lenz searches the public record when it runs the check, so a published change is in scope; every source and the check itself are dated. Where superseded guidance is still online beside the new rule, a deep check returns both sides and the reasoning. Put the tax period or the date in the statement to anchor the check.

What you send through the API or a connector is private to your account, and each adviser’s checks are their own. The example on this page is a memo paragraph Lenz wrote and published. Anthropic, Google, OpenAI, Perplexity and xAI, the AI model providers Lenz uses, do not train their models on what Lenz sends them. By default Lenz keeps your memos and their results while your account exists; on Pro and Scale you can set a shorter period, including zero, where what you submit is never written to Lenz’s database. The model providers keep what their own terms say, listed on the subprocessors page.

Every check you run is private by default, on the website, through the API and in the Claude connector. On a paid Pro or Scale plan you can set zero retention: what you submit is never written to our database, a result stays readable from our cache for 24 hours, and the model and search providers keep what their own terms say.

Check the memo before it reaches the client.