AI chatbot for law-firm intake

The safest intake chatbot does less than most demos promise. It identifies itself, gathers limited facts, checks basic routing criteria, schedules the right next step, and stops before legal judgment.

That narrower job is deliberate. DATIK will not configure a public intake bot to improvise legal advice.

The firm remains responsible for confidentiality, supervision, communication, competence, and reasonable fees.

Start with administrative intake

Good first tasks include identifying the matter type, jurisdiction, opposing parties for a conflict-check process, deadlines stated by the visitor, preferred contact method, and availability for a consultation.

The bot should not decide that a person has a valid claim, predict an outcome, or provide individualized legal advice. It should avoid creating the impression that an attorney-client relationship already exists.

Disclose that it is AI

Tell the visitor at the beginning. Use plain language. The disclosure should also explain that the chat is not legal advice, does not create an attorney-client relationship, and may be reviewed by the firm.

Do not bury the disclosure after the visitor submits facts. Florida Bar Ethics Opinion 24-1 says a generative AI chatbot communicating with clients or third parties must state that it is an AI program, not a lawyer or law-firm employee. The opinion is advisory rather than binding, and other jurisdictions may use different rules.

Limit confidential information

Ask for the minimum needed to route the inquiry. Warn the visitor not to send detailed confidential information before the firm completes its intake and conflict process.

The ABA's Formal Opinion 512 says lawyers using generative AI must consider competence, confidentiality, communication, and fees. Review the vendor's terms, retention, security, access controls, and model-training provisions before any client information enters the system.

Keep conflicts outside the public bot

The bot may collect names needed for a conflict check, but the firm's established system should perform the check. Do not expose match details or let the bot declare that no conflict exists.

Route urgent deadlines to a person with a clear warning that submission does not guarantee review or representation.

Create a narrow knowledge base

Use approved information about practice areas, office details, consultation process, public fees, and general educational material. Exclude internal work product and matter files unless the deployment has a separately reviewed security design.

Every answer should trace back to an approved source. Unknown questions move to a person.

Test the risky conversations

Test a represented party, an adverse party, a minor, an urgent deadline, a request for legal advice, a request to upload evidence, a threat of harm, and a visitor outside the firm's jurisdiction. Counsel should approve the response and handoff for each case.

Measure intake quality

Track complete inquiries, qualified consultation requests, conflict-check failures, urgent handoffs, abandoned chats, unanswered questions, and incorrect or overbroad answers. Review transcripts under the firm's retention and access policy.

The DATIK Lead Chatbot can be limited to approved material and human handoff. A law-firm deployment still requires the firm's ethics and security review.

Sources and review limits

• ABA overview of Formal Opinion 512
• Florida Bar Ethics Opinion 24-1
• NIST Generative AI Risk Management Profile

Sources were checked on September 26, 2026. This article is general information, not legal or ethics advice. Rules differ by jurisdiction and facts. The firm should consult its responsible lawyers, insurer, state bar guidance, and security professionals.