Use case · Law

Legal AI sounds right. A lawyer checks whether it is.

Legal AI writes fluent summaries, citations and redlines that still miss the clause that matters. Lasker supplies review by admitted lawyers: evals for legal-AI teams, and review inside the firm when client work cannot leave.

For teams building contract review, legal research and discovery tools, and for firms deploying them.

Admitted lawyersChecked with the admitting body
Every citation checkedExists and says what is claimed
Jurisdiction recordedOn every record
Inside the firm if neededOnly findings leave
// Where models break

Where legal AI gets sent back.

A fluent answer is the easy part. These are the places a partner or a court finds the problem.

Citations

Cases and statutes that do not exist, or do not say what the summary claims. Courts have sanctioned lawyers who filed AI-invented citations.

Cited case not found in any reporter.

Contract review

The clause that matters is often the exception to the clause the model summarised: carve-outs, caps, survival.

Cap in 9.2 excludes indemnity claims.

Jurisdiction

Governing law and state rules change the answer. California, for one, generally voids employee non-competes.

Non-compete analysis ignores California law.

Privilege in discovery

A privileged document produced by mistake can put protection at risk. Review has to catch it before production.

Email to outside counsel marked responsive, not privileged.

Deadlines

Response periods depend on the court’s rules, the method of service and holidays.

Deadline counted from the wrong trigger date.

Drafting consistency

Defined terms, cross-references and conforming changes across a long document.

“Effective Date” defined twice.

Lawyers must keep client information confidential when they use AI, and generally need the client’s informed consent to share it. So the work that teaches legal AI the most rarely leaves the firm.

Benchmarks test whether a model knows the law. A firm needs to know whether it handled this clause, in this deal, under this law.

A lawyer stays responsible for what goes out under their name, so every AI draft still needs a lawyer’s review.

// What a production tool needs

Six needs, and where the data comes from.

For teams whose legal AI produces work a lawyer has to stand behind.

Know what a lawyer rejects

Review notes on real outputs: what was wrong, why, and the fix.

Source: admitted lawyers reviewing your tool’s outputs

Check every citation

Each authority found, read and matched to the claim it supports.

Source: citation checks by lawyers

Know the jurisdiction

Answers checked by a lawyer admitted where the law applies.

Source: reviewers matched to the jurisdiction

Spot privilege

Privilege calls on realistic document sets, with the reason.

Source: reviewers with discovery experience

Pick the better draft

Two drafts for one task, ranked by a lawyer, with the reason.

Source: preference data from reviewers

Work where the matter is

Review on live matters, inside the firm’s systems.

Source: on-site review; only findings leave

// What we deliver

Six deliverables, scoped per pilot.

Evals for legal-AI teams use your tasks and documents without client data. Review on live matters happens inside the firm.

DatasetWhat is in a recordUsed forFormat
Expert evalsTool output scored against a written rubric by an admitted lawyerEvals, release gatesScores + rationale, JSONL
Citation checksEvery authority in a draft: found, accurate, supports the claimHallucination testsJSONL per citation
Redline reviewClause-level findings on a contract review, with the correct positionContract review trainingJSONL + marked-up document
Preference dataTwo drafts for one task, ranked, with the reasonReward modelsJSONL
Workflow tracesA lawyer doing the task step by step on a public or synthetic matterDemonstrations, fine-tuningJSONL + source documents
Review inside the firmLawyers review the tool on live matters in the firm’s environmentDeployment evalsFindings the firm approves
// Sample record

What records look like.

A lawyer’s eval of a contract summary, and a citation check on a research memo.

  • scoresPer rubric criterion, so you can train on what failed
  • verdictCorrect, needs edits or incorrect, as a reviewing lawyer would put it
  • issuesEach citation problem with its type
  • jurisdictionThe law applied, recorded on every record
  • rightsSource and whether any client data was involved

Illustrative. Field names are our draft schema, and we adapt it to your pipeline in the pilot.

lg-0233.jsonIllustrative
{
  "record_id": "lg-0233",
  "task": "Summarize limitation of liability, MSA",
  "jurisdiction": "New York",
  "reviewer": {
    "credential": "NY bar",
    "verified": true
  },
  "scores": {
    "accuracy": 2,
    "authority": 5,
    "completeness": 3
  },
  "finding": "Cap in 9.2 excludes indemnity claims.",
  "verdict": "incorrect",
  "rights": {
    "source": "vendor task, no client data",
    "personal_data_removed": true
  }
}
ct-0091.jsonIllustrative
{
  "record_id": "ct-0091",
  "draft": "research memo, 14 citations",
  "jurisdiction": "California",
  "checked": 14,
  "issues": [
    { "cite": 6, "type": "not_found" },
    { "cite": 11, "type": "misstated_holding" }
  ],
  "reviewer": {
    "credential": "California bar",
    "verified": true
  }
}
// Expert review

How lawyers score legal AI.

The questions a reviewing partner asks, written down before scoring so two reviewers can agree.

DimensionWhat the reviewer checksEvidence per finding
Legal accuracyThe statement of the law and the facts is rightRule or clause cited
AuthorityEvery citation exists and supports the pointCitation-by-citation result
JurisdictionThe right law is appliedGoverning law noted
CompletenessThe issues a lawyer would raise are raisedMissing issues listed
UsabilityCould go out after light editsEdits needed

Reviewers are admitted in the jurisdiction they score, checked with the admitting body. Items with disagreement go to a second reviewer.

// Pilot

A pilot on one task.

Pick the task where your tool is weakest: contract review, research or discovery. The pilot answers whether this review moves it.

Scope

One task, one jurisdiction and the failure modes you care about.

Week one

Sample

A set of outputs scored by admitted lawyers against a rubric you approve.

You test it in your pipeline

Decide

Full set, ongoing review or stop. Price is set after the sample.

No commitment before

We are recruiting admitted lawyers now. No legal dataset or engagement has been delivered yet.
// FAQ

Questions labs ask.

Data handling and our commitments are in the Trust center.

Who are the reviewers?

Lawyers admitted in a US state, solicitors in England and Wales and barristers, checked with the admitting body and matched to the jurisdiction of the task.

Do you need our clients’ data?

Not for evals: they use your tasks and documents with no client data. Review on live matters happens inside the firm’s environment, and only the findings the firm approves leave.

Is this legal advice?

No. Reviewers evaluate AI output for your team. It is not legal advice to you or to your clients.

Which jurisdictions?

US states and England and Wales first. The law applied is recorded on every record.

Can a model trained on it ship commercially?

The license scope is written per dataset: training, evaluation or both, and whether a model trained on it can be used commercially.

How is it priced?

Per batch of tasks or per engagement, after the sample. We do not publish ranges until we have closed deals to base them on.

Run a firm? Have lawyers review your AI tools inside the firm, on your systems. Only the findings you approve leave.

Email as@lasker.ai →
// Scope a pilot

Tell us where your legal AI fails.

Send the task, the jurisdiction and the volume. We reply with a rubric draft and what a pilot would include.