When a Prospect Arrives From ChatGPT: Intake for AI-Informed Legal Inquiries

By Michael Brunman, Co-Founder & CEO of Clerx, Former Litigator

The direct answer

AI-informed intake is the process of receiving a prospective client who has already consulted an AI tool or answer engine, clarifying what they want from the firm, and routing them to the appropriate professional review. It should preserve the person’s question without endorsing an AI-generated legal conclusion.

Clio’s 2026 Legal Trends Report reports that 46% of consumers would consult AI before contacting a lawyer. That is stated willingness, not a measured share of your leads or proof that ChatGPT referred them to your firm.

For a lean practice, the response should be practical: identify the service requested, ask approved screening questions, explain the next step, and leave legal assessment to the attorney. Clerx AI Receptionist can support that configured operational conversation.

Ask what brought the person to the firm

A person may say they found the practice through ChatGPT, an AI summary in search, or a friend who used an AI tool. Those are different source descriptions. Record what they actually report.

The lead-attribution guide helps avoid false precision. A self-reported source is useful evidence but is not a verified reconstruction of every step in the person’s research.

Do not ask for account access or a complete chat history merely to attribute the inquiry. The first task is serving the prospective client, not making them prove the marketing journey.

Separate the question from the asserted answer

Someone may say, “The AI told me I qualify,” or “It said I only need one document.” Intake should record the request without validating the claim.

An illustrative response is: “The attorney can review that with you. I can help identify the right consultation.” The firm should approve the wording and screening questions.

The guide to AI and human intake responsibilities explains the broader division of work. An AI receptionist should not debate another model’s legal answer or replace attorney review with a second automated opinion.

Confirm the firm’s actual services

An answer engine may describe a service the firm does not offer, use an outdated fee, or imply a particular attorney is available. Use the firm’s current approved information to explain what can actually be booked.

The intake-question framework helps establish fit without a long interrogation. Ask the broad service category, relevant location where needed, existing-client status, and any reported timing concern.

If the person’s request does not fit, provide the firm-approved response. Do not stretch the service description simply because the person arrived with a confident AI recommendation.

Correct operational misinformation carefully

The receptionist can explain the firm’s actual consultation fee, office hours, supported languages, or appointment format when those facts are approved. It should not resolve legal misinformation about rights, deadlines, or likely outcomes.

The distinction matters. “Our initial meeting is by video” is an operational correction. “The AI’s eligibility analysis is wrong” requires legal judgment beyond the receptionist’s role.

The conflict-review guide reinforces another boundary: an AI recommendation does not establish that the firm can accept the matter or that conflicts have been cleared.

Give the attorney a useful summary

A helpful record describes the person’s request, broad context, what they report having been told, and what remains unresolved. Keep the source attribution visible.

For firms using MyCase or Clio, verify where the summary reaches the team. Do not copy an entire AI conversation into a record by default when a concise account of the question is sufficient.

The security checklist helps define access and handling. Staff should use approved routes if the attorney later requests supporting material.

Connect the inquiry to a real consultation

A prospective client may believe that AI research has already completed the professional assessment. Explain what the consultation actually offers, including fee, duration, format, and any review required before booking.

Clerx scheduling and payments can support configured consultation workflows. The relevant benefit is a clear route from inquiry to the correct appointment, not a promise that an AI-sourced inquiry will convert better.

If the firm uses Clio’s own paid-booking route, keep Clio Scheduler and Clio Payments accurately distinguished from other ecosystems. A search referral should not become an excuse to blend unrelated product capabilities.

An immigration inquiry example

Consider a fictional Massachusetts solo immigration lawyer. A caller says an AI tool recommended a particular application and wants the office to “file it this week.” The firm requires a consultation before deciding what services it can provide.

The receptionist should capture the broad question and caller-reported time concern, explain the consultation path, and avoid affirming eligibility or a filing commitment. The attorney receives the reported AI recommendation as context, not an approved case strategy.

If the caller asks whether the AI answer is correct, the approved response should keep legal review with the attorney. If they ask the consultation fee, the receptionist can provide the current approved amount.

Now test the same inquiry through website chat. The boundary should remain consistent even when the person pastes a confident paragraph into the conversation. A written assertion does not become verified merely because it looks detailed.

Use attribution to improve content responsibly

After enough inquiries, the firm may see recurring questions or inaccurate descriptions of its services. Review these patterns without treating a small sample as measured search demand.

The guide to how AI search understands law-firm websites provides broader visibility context. This article concerns the next step after someone contacts the office.

The firm can clarify its public service descriptions and consultation pages where appropriate. Avoid publishing a person’s private AI conversation or legal issue as a marketing example without authorization.

Test and measure the new entry point

Use fictional inquiries involving an incorrect fee, unsupported service claim, AI-generated eligibility conclusion, and person who does not remember which tool they used. Inspect whether the operational corrections and legal boundaries remain clear.

Track self-reported AI sources, requested services, booking outcomes, and staff clarification work using stable definitions. These are proposed original-data opportunities. They do not establish causal conversion effects or a universal superiority of AI-originated leads.

The best result is a prospective client who understands what the firm can do next and a lawyer who receives a dependable account of the question.

Keep the source conversation proportionate

A person may be embarrassed about using an AI tool or may not consider it relevant. Make source questions optional where the firm’s process allows, and do not let attribution delay service.

The receptionist can ask how the person found the practice and preserve the response. If they say “online,” do not infer ChatGPT. If they say an AI tool suggested the firm, do not assume that tool originated the underlying legal question.

That discipline matters when the owner later reviews marketing results. A handful of voluntary source reports can reveal a useful pattern, but cannot establish market share, a conversion advantage, or a complete acquisition path.

Prepare for confident but incomplete narratives

AI-generated explanations can sound organized while omitting facts the attorney needs. Use the same approved screen applied to other inquiries rather than treating polished wording as evidence of fit.

A person may present a specific application, claim, or document as the answer. Ask what they want the firm to help with and record the assertion as reported. Do not rewrite the inquiry into the AI’s chosen legal category unless the firm’s reviewer approves that classification.

Include a branch for uncertainty. Someone may arrive convinced they need a service the firm does not offer, while their underlying question still needs professional review. The receptionist should use the firm’s routing instructions rather than attempting a legal reclassification.

Improve public descriptions from repeated operational questions

If several callers arrive with the same wrong consultation fee or service expectation, check the firm’s public pages and approved intake instructions. Correcting the source material may be more useful than training staff to explain the same mismatch repeatedly.

Keep the evidence specific. Record what callers reported and what the firm verified. Do not claim that a particular answer engine invented a statement unless the underlying output was actually reviewed.

The firm can publish clearer explanations of its services, appointment scope, and what intake can do. Any content about legal options should receive the appropriate attorney review. The objective is an understandable route to professional help, not a promise that a site update will control every AI-generated answer.

During a pilot, compare staff clarification work for AI-informed and other inquiries using consistent definitions. Differences may reflect practice mix, marketing source, or sample size. Treat the findings as descriptive until there is enough evidence to support a stronger conclusion. This gives the owner useful operating information while avoiding unsupported claims about the quality of AI-originated leads.

Keep the eventual content response separate from the intake record

Recurring AI-informed questions may suggest a useful educational article, but the public explanation should be developed independently of any individual inquiry. Use aggregated themes or fictional examples and obtain the appropriate professional review.

A person asking about their own situation has contacted the firm for help, not supplied a marketing case study. Do not turn a transcript or pasted AI output into promotional content merely because it illustrates an interesting mistake.

When reviewing patterns, separate service-description problems from legal questions. An outdated office-hour statement can be corrected operationally. An article explaining a legal issue needs attorney review and should not imply an individualized answer for future callers.

The firm can also refine intake instructions based on observed confusion. If people repeatedly mistake an introductory meeting for a completed legal assessment, make the consultation explanation clearer in every relevant channel.

Measure whether those changes reduce clarification work or booking misunderstandings. Avoid claiming that they improve answer-engine rankings without evidence. The immediate, testable benefit is a better-informed conversation when a prospective client reaches the office, regardless of where their research began.

Frequently asked questions

What is an AI-informed inquiry?

An inquiry from someone who has already used an AI tool or answer engine for research. Their reported research is context, not a legal determination.

Does Clio’s 46% figure measure referrals?

No. It describes stated willingness to consult AI before contacting a lawyer, not the percentage of leads generated by AI tools.

Should intake validate a ChatGPT legal answer?

No. Preserve the question and route legal assessment to the attorney.

Can the receptionist correct an outdated fee?

It can provide the firm’s current approved operational information. Legal conclusions require professional review.

Should the firm request the whole AI chat?

Not by default. Collect only what the approved intake process needs and use an authorized route for further material.

How should the source be recorded?

Preserve the person’s own description and label it as self-reported. Avoid inventing a precise attribution path.

Can AI-sourced inquiries book directly?

Only under the firm’s screening and appointment rules. The source does not bypass review or acceptance requirements.

What should the firm measure?

Self-reported sources, fit, booking completion, and clarification work, with small samples and attribution limits made explicit.

Book a demo with Clerx today.

About the author

Michael Brunman is Co-Founder & CEO of Clerx, helping law firms improve client intake and communication. A former litigator, software engineer, and McKinsey consultant with an MBA from Harvard Business School, he brings legal, technical, and business experience to the challenges of running a law firm.