11/16/2026
7 Plain-Language AI Receptionist Disclosures for Law Firm Intake
Explain who is answering, what the assistant can do, and how to reach the team without overwhelming the caller.

By Michael Brunman, Co-Founder & CEO of Clerx, Former Litigator
An AI receptionist disclosure tells callers what they are speaking with, what it can help them do, and how to reach the firm's team. A useful disclosure is short enough to understand during a call and specific enough to prevent someone from mistaking intake assistance for legal advice.
For a solo lawyer, the challenge is practical. A long warning at the beginning can overwhelm an already anxious caller. A vague introduction can leave that person assuming an attorney or employee is making decisions. The better approach is a clear opening followed by additional explanations when they become relevant.
This guide provides seven operational wording patterns. They are starting points for a firm's approved script, not a statement that any particular sentence satisfies every jurisdiction's disclosure, recording, privacy, or professional responsibility requirements.
Why disclosure belongs in the intake design
Clio's 2026 Legal Trends Report announcement, published October 26, 2026, reports substantial consumer interest in transparency about lawyers' AI use. That supports a practical lesson: do not make the caller discover the technology by asking an uncomfortable question halfway through the conversation. It does not establish that one particular receptionist greeting improves conversion.
The ABA's introduction to Formal Opinion 512 identifies competence, confidentiality, communication, and informed consent among relevant duties. Its application depends on circumstances and applicable rules. A greeting alone cannot address those obligations.
Treat transparency as part of the broader AI receptionist evaluation. The firm still needs approved questions, a defined scope, and a person responsible for reviewing exceptions.
1. Identify the assistant without turning the greeting into a lecture
Suggested wording: "Thank you for calling the firm. I'm its AI intake assistant. I can take your information and help with the next step."
Use the actual firm name and an introduction consistent with the configured service. Avoid job titles that suggest the assistant is an attorney, paralegal, or human employee. If the assistant has a name, the name should accompany the explanation rather than substitute for it.
An assistant called Donna can still introduce itself as an AI intake assistant. The caller should not have to infer its nature from its voice or name. Read the introduction aloud, including the firm name. A sentence that looks short in a document can become cumbersome when spoken.
For Clerx's inbound AI receptionist, the introduction should fit the firm's configured call workflow. Do not use a script that promises actions that have not been enabled.
2. Explain the useful job the assistant can perform
Suggested wording: "I can collect a few details about your inquiry and help arrange the appropriate next step."
People call with an objective. They want a consultation, a callback, a document-delivery instruction, or help contacting an existing legal team. Explain the relevant assistance, then ask a simple question.
If direct booking is available for that inquiry type, the script can describe it. If the firm requires review first, describe a callback request instead. Saying "I can book you" and later withdrawing the offer creates avoidable confusion.
Keep this explanation aligned with the firm's scheduling configuration. A calendar connection does not mean every matter type, attorney, appointment length, or location should be offered to every caller.
3. Separate intake questions from legal advice
Suggested wording: "I can help with intake and scheduling. An attorney would need to review legal questions and advice."
Use this boundary when a caller asks what to do, whether a case will succeed, or what a document means. The assistant can capture the question in the caller's own words without answering it substantively.
For example, an immigration caller may ask whether travel is safe. The intake step is to record that travel is the reason for the request and route it under the firm's instructions. The assistant should not evaluate immigration consequences. Similarly, an estate-planning caller asking whether an existing will is valid needs attorney review, not a confident conversational answer.
The firm's intake forms and question design should reinforce the same boundary. A field labeled "eligible" can imply a legal conclusion even when the underlying conversation was cautious.
4. Give an honest route to a person
Suggested wording: "If you prefer to speak with the team, I can follow the firm's contact instructions."
Replace that wording with the actual available option: an attempted transfer, a callback request, or another approved route. Do not promise an immediate staffed response if no person is available.
An overnight caller can reasonably choose a callback after hearing when the office expects to respond. That is different from being told someone will take over immediately and remaining on hold. The assistant should describe the actual next step before trying it.
Build this into the caller-asks-for-a-human policy. Human fallback means a workflow involving the firm's people or its separately arranged service. It should not imply Clerx provides a staffed human answering team.
5. Address recording and information handling separately
Suggested operational pattern: provide the firm's approved recording and privacy language at the point required by its policy and applicable law, then explain any available alternative.
There is no universal recording sentence supplied here. Requirements can depend on the participants, jurisdictions, technology, and use. Have the responsible lawyer approve the actual language and sequence. Do not assume that announcing AI also announces recording, or that permission to leave a message authorizes every later use of that message.
Review the security and privacy buying questions alongside the script. Ask who can access records, how the account is configured, and what retention terms apply. Do not promise deletion or a specific security control merely because it sounds reassuring.
6. Make written-channel explanations visible where the conversation begins
Suggested wording: "You're chatting with the firm's AI intake assistant. Please share a brief description so we can help with the next step."
For website chat, place the explanation where the visitor sees it before relying on a substantive response. A tiny footer on a separate page is a weak substitute for a visible conversation introduction.
For text messaging, use channel-appropriate identification and the firm's approved messaging policies. Written conversations can be reread, forwarded, or seen on a shared device. Ask for only the information needed for the immediate operational purpose.
The wording need not be identical across calls and messages. The identity, scope, and human route should remain consistent. A caller who moves from phone to chat should not encounter contradictory descriptions of who is responding.
7. Explain what happens after information is collected
Suggested wording: "I'll record your request for the team. They will review it and follow the firm's next-step process."
Be more specific when the firm has a reliable process: an appointment confirmation, a named review queue, or a callback window it can actually meet. Avoid saying "your case has been accepted" when the event is only a lead record being created.
If an integration connects intake to practice management, callers do not need a technical explanation of the connection. They do need an accurate understanding of the outcome. "Your request has been recorded" is different from "an attorney has reviewed your matter."
End with the next step, not an unexplained assurance that everything is handled.
Test the explanation in realistic conversations
Use fictional test calls before changing a live script. Include a caller who interrupts the introduction, one who asks directly whether it is AI, one who requests a person, and one who starts describing a sensitive legal issue immediately.
Listen for comprehension rather than merely checking that a sentence was spoken. Does the assistant answer the identity question plainly? Does it repeat a long disclaimer unnecessarily? Does it preserve the boundary when the caller asks the same legal question in different words?
Include the firm's accessibility review. Speech speed, repeated questions, noisy connections, and language switching can change whether the introduction is understandable. A successful quiet test call does not prove the same experience for every caller.
An office manager can record three observations after each test: what the caller understood, what action was taken, and what remained unresolved. Correct the misunderstanding that matters most before polishing minor phrasing.
Keep the disclosure short and the operating policy complete
A concise greeting works only when the process behind it is defined. Maintain a short approved script plus a separate internal explanation of recording, privacy, escalation, booking eligibility, and attorney review.
For a one-to-five-attorney firm, one person can own changes, with the responsible lawyer approving legal and professional responsibility language. Avoid letting several employees edit separate versions across phone, chat, and messages without coordination.
Do not measure success only by how quickly the assistant gets past its introduction. Track requests for clarification, unresolved human requests, incorrect booking expectations, and staff corrections. These observations can reveal whether the explanation is helping callers understand the service. They are proposed operating measures, not published Clerx performance results.
Keep exceptions visible to the next person
A caller may understand the explanation and still need help the assistant cannot provide. Record that unresolved need alongside the next step, rather than hiding it behind a successful-call label. For example, a person who requests an attorney and accepts a callback has not received legal advice or completed a consultation. The staff member reviewing the record should be able to distinguish those outcomes immediately. This is a proposed review practice, not a claim about an automatic Clerx field or reporting feature.
Frequently asked questions
Should an AI receptionist identify itself?
A plain identification is a sensible transparency practice. The firm must separately determine what disclosures its applicable rules and configuration require. The script should never falsely describe the assistant as a human or attorney.
Does an AI disclosure also cover call recording?
Do not assume so. AI identity and recording are different subjects. Use separately approved language and timing for the firm's actual recording practices and legal obligations.
Can the receptionist use a personal name?
Yes, a name can make the interaction easier to follow. Pair it with an accurate explanation of the assistant's role so the name does not create a misleading impression.
Must the opening list every limitation?
The opening should communicate identity and purpose clearly. Additional explanations can appear when relevant, subject to the firm's approved legal requirements. A long opening is not automatically a more effective disclosure.
What if a caller wants a person immediately?
Follow the configured human-contact route and describe its actual availability. If only a callback request is possible, say so instead of promising an immediate transfer.
Can the assistant assess whether a case is strong?
It can collect facts and apply approved operational screening criteria. Legal assessment and advice remain with attorneys. Avoid presenting a qualification label as a legal conclusion.
Should the same wording appear in website chat?
Keep the identity and scope consistent, while adapting the wording to a visible written conversation. The channel should not contradict what the firm tells callers by phone.
How often should the firm review its script?
Review it when channels, recording practices, booking rules, or human routes change, and after recurring misunderstandings. Keep one approved version rather than relying on informal staff edits.
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.
Recommended Posts
11/21/2026
Clerx vs Specialty Answering Service: Legal AI Intake, Live Reception, and Hybrid Options
Compare the actual intake workflow, human fallback, legal-software handoff, and complete cost rather than reducing this choice to AI versus people.
11/20/2026
Lawmatics Form Submission vs AI Field Updates: Why an Intake Automation May Not Run
Information can arrive successfully while the event your next workflow expects never occurs.