· 8 min read

Legal intake AI receptionist: what to capture before the consult

A practical legal intake call flow for solo and small law firms. What an AI receptionist can safely collect, what it must avoid, and how to route qualified matters without giving legal advice.

A refined law office intake desk with a phone, case folder, and abstract AI matter-intake dashboard

Legal intake is one of the best use cases for an AI receptionist, but only if the boundary is clear.

The AI should not give legal advice.

It should not evaluate case merit.

It should not tell the caller what their rights are.

It should collect structured facts, route by your rules, schedule when appropriate, and protect the attorney’s time from unqualified calls.

That is still a big win.

The first intake call should answer:

  • Who is calling?
  • What type of matter is this?
  • Where did it happen?
  • When did it happen?
  • Is there a deadline?
  • Who are the opposing parties?
  • Is the caller a fit for the firm’s practice areas?
  • Is the matter urgent?
  • Should this become a consult, referral, decline, or human review?

Most of that is administrative. It does not require legal analysis. It does require consistency.

That is why AI can help.

The first 30 seconds

The AI should open with a disclaimer-like boundary in plain language:

“I can collect intake details for the firm and help route your request. I cannot provide legal advice or tell you whether you have a case.”

Then it should ask:

“Are you a new caller, an existing client, opposing counsel, court staff, or calling about something else?”

This matters because law firm phones mix new leads with existing client calls, courts, vendors, opposing counsel, spam, and referral partners. Treating every call as a new lead creates chaos.

New-client intake fields

For a new prospective client, capture:

  • Full name
  • Callback number
  • Email
  • Best time to reach them
  • Practice area in caller’s words
  • Location/jurisdiction
  • Date of incident or issue
  • Any known deadline or hearing date
  • Opposing party names
  • Employer, spouse, business, insurer, landlord, or other relevant parties depending on practice area
  • Whether they have already hired a lawyer
  • Whether they have upcoming court dates
  • How they found the firm
  • Permission to text/email if your process uses it

The AI should ask only what the firm actually needs. A personal injury firm, family law firm, immigration firm, criminal defense firm, and estate practice should not use the same script.

Conflict-check information

Conflict checks are one of the biggest reasons legal intake needs structure.

At minimum, the AI should capture opposing party names before a consult is confirmed if that is part of your policy.

For example:

  • For family law: spouse or other party name
  • For employment: employer name
  • For personal injury: defendant, business, insurer, or employer if known
  • For business disputes: company names and key individuals
  • For landlord-tenant: landlord, property manager, or tenant name

Some legal AI platforms market real-time conflict checks against tools like Clio, Filevine, Lawmatics, MyCase, and PracticePanther. That can be useful. But the key is policy. The AI should not say “conflict cleared” unless the firm has approved that specific automated workflow.

Safer language is:

“I am collecting names for the firm’s conflict review.”

Urgency classification

Legal urgency is not always obvious. A caller may sound calm but have a deadline tomorrow. Another caller may sound panicked about something routine.

The AI should ask:

  • Is there a court date?
  • Is there a response deadline?
  • Has the caller been arrested or served?
  • Is there a protective order, eviction date, deportation deadline, wage garnishment, or hearing?
  • Is anyone in immediate danger?

Then route according to the firm’s policy.

Again: no legal advice. The AI should not say whether a deadline applies. It should ask whether the caller knows of one and alert the human team.

Practice-area routing

Most small firms lose time on non-fit calls.

The AI can help by routing:

  • Fits practice area and location: schedule consult or send to intake team
  • Practice-area mismatch: polite referral language or callback queue
  • Existing client: route to staff
  • Emergency/urgent: alert human
  • Spam/vendor: filter

This is where a legal AI receptionist earns its keep. It does not need to be a lawyer. It needs to stop every call from landing on the attorney’s desk as “please call back.”

What the handoff should look like

A useful legal intake summary:

“New family law intake. Caller: Amanda R. Callback: 555-0144. Email provided. Jurisdiction: Harris County, TX. Issue: divorce and custody. Opposing party: Daniel R. Court date: none known. Has not hired counsel. Found firm via Google. Wants consult this week. No immediate safety issue reported.”

That gives the attorney or paralegal a real starting point.

A bad summary:

“Caller wants divorce help.”

The whole point of intake automation is to remove the second call where staff has to ask everything again.

Confidentiality and AI

Law firms need to think carefully about confidentiality and supervision. ABA Formal Opinion 512 discusses lawyers’ duties when using generative AI, including competence, confidentiality, communication, and supervision.

For intake, the practical implications are:

  • Use approved tools and accounts.
  • Know where call data goes.
  • Avoid sending confidential details into consumer AI tools.
  • Keep humans responsible for legal judgment.
  • Make sure scripts and outputs are reviewed.
  • Document what the AI is allowed to do.

Do not treat a public chatbot as legal intake infrastructure. That is not the same thing as a controlled phone workflow with access rules, storage decisions, and firm-approved language.

What not to automate

Do not automate legal advice.

Do not automate case valuation.

Do not automate “you have a case” or “you do not have a case.”

Do not automate privilege explanations unless the firm has approved exact language.

Do not let AI respond to substantive legal questions beyond routing and intake.

The safe deployment is administrative: capture, classify, schedule, summarize, and route.

Where AI fits best

AI legal intake is strongest for:

  • Solo and small firms with consumer-facing inbound
  • Personal injury
  • Family law
  • Criminal defense
  • Immigration
  • Estate planning
  • Employment
  • Landlord-tenant
  • Bankruptcy

It is weaker for practices where every first call requires nuanced legal analysis, complex conflict review, or high-stakes confidentiality decisions before any facts are captured.

If the first five minutes are mostly administrative, AI can help. If the first five minutes are legal judgment, keep a human there.

The deployment I would build

For a solo or small firm, I would build:

  1. AI answers new calls and separates new lead, existing client, court/opposing counsel, vendor, and spam.
  2. It collects practice-area-specific intake fields.
  3. It captures opposing party names for conflict review.
  4. It flags urgent deadline and court-date language.
  5. It schedules consults only inside firm-approved rules.
  6. It sends a structured intake summary to the attorney or paralegal.
  7. It never gives legal advice.

That is the practical shape behind Legal Intake AI Receptionist.

If you want this mapped to your firm, send your current intake questions through the free workflow audit. I will tell you where AI fits and where it should stay out.

Sources reviewed

Related operator notes

Keep reading

No-pressure first step

Not sure which one fits?
Get a free 20-min audit.

Bring one workflow you'd want automated. I'll tell you which deployment fits — and which doesn't — in twenty minutes. No pitch deck, no follow-up sequence. Useful even if you don't buy.

  • A real plan, not a sales call

    Which surface (Telegram, Discord, Slack, phone) fits your team, and which one doesn't.

  • Honest "don't buy this" if it applies

    If a $99/month SaaS solves it, I'll tell you which one and how.

  • A timeline + price range

    When I could deploy, what it'd cost, and what you'd own at the end.