We built DocketReply around the rules, not in spite of them.
ABA Op. 512. Florida 24-1. Texas 705. New York 2025-6. California COPRAC. Every state bar that has weighed in on generative AI has said roughly the same thing: understand the tool, supervise the output, protect the client, and tell the truth. DocketReply is built so that doing the right thing is the default.
14-day free trialNo credit cardCancel anytime
The bar opinions on AI that you actually need to know.
We can't give you legal advice, and you should review the actual opinions with your own counsel and malpractice carrier. But here's the short version of where the law currently sits.
American Bar Association
Formal Opinion 512 · July 2024
Lawyers using generative AI must understand the technology, supervise outputs, protect confidentiality, communicate use to clients where required, charge fees ethically, and verify accuracy before relying on output.
Florida
Bar Ethics Opinion 24-1 · January 2024
AI chatbots interacting with the public must disclose they are AI, not a lawyer. Confidentiality requires informed consent before sending client data to third-party AI. Cost passthroughs must be in writing and at actual cost.
Texas
State Bar Opinion 705 · February 2025
Lawyers must independently verify AI output, protect confidentiality (informed consent for third-party AI), bill ethically (no double-charging for time saved), supervise output, and understand the tool they use.
New York
NYSBA AI Task Force / Op. 2025-6 · 2025
Phased adoption framework with focus on AI recording/transcription of client meetings. Confidentiality and consent are emphasized for any AI tool that touches client communications.
California
COPRAC Practical Guidance · November 2023
Practical Guidance on the Use of Generative AI in the Practice of Law: competence, confidentiality, accuracy, candor to tribunal, communication, and supervision are all affected duties.
Six things that are true the moment you create your account.
No setup steps, no opt-in. The defaults are the safe ones.
We don't draft legal arguments.
DocketReply only drafts the first reply — confirming receipt, expressing empathy, warning the client not to give recorded statements, and offering a call. No case law, no claim valuation, no statute of limitations analysis. Far lower risk surface than tools that draft motions or briefs.
Bar-compliance mode is the default.
Every new account starts with bar-compliance mode on. Auto-send is disabled. Each reply requires your manual approval before it leaves your inbox. You opt out when you're ready, not before.
State-bar disclaimer is appended automatically.
Pick your jurisdiction in settings. We append a vetted disclaimer aligned with your state bar's opinion to every outbound message. Plain text and HTML, in the message body — not buried in a footer image.
Two-pass safety review before any auto-send.
If you do enable auto-send, every draft passes a regex safety gate plus a Claude Haiku reviewer that reads the draft as a state-bar ethics auditor. Anything that promises an outcome, implies an attorney-client relationship, or quotes a fee percentage drops back to your review queue.
Audit trail for your malpractice carrier.
Every draft generated, every approval, every send, every settings change — logged with timestamp, actor, and IP. Exportable as CSV (Firm tier and above) so your carrier can see exactly what the AI did and what you reviewed.
We don't train on your client data.
Your intake submissions are sent to Anthropic and Google for completion only. Both providers' enterprise APIs do not train on customer data by default. Nothing leaves your tenant for model training.
That's the AI hallucination case
that scares every PI lawyer.
In March 2026 the Sixth Circuit fined two attorneys $15,000 each (plus opposing fees) for filing a brief with twenty-four fabricated citations generated by AI they didn't verify. It's the most-cited cautionary tale of the year.
DocketReply doesn't draft briefs. It doesn't draft motions. It doesn't cite cases. It writes a first-touch reply to a prospective client — confirming you got their inquiry, asking a follow-up question, warning them about insurance adjusters, and offering a free call. The lowest-risk surface there is for a PI firm to introduce AI into. Whiting could not happen here.
14-day free trialNo credit cardCancel anytime