The Claude playbook built for small Florida law firms.
Not for BigLaw. Not for tech companies. For you.
15 ready-to-adapt prompts across seven workflow stages. Full ethics framework. A governance policy template to adapt for your firm.
Primary references: Anthropic commercial data handling for Claude and Florida Bar Ethics Opinion 24-1.
Download the Playbook →Generic AI advice for lawyers was written for firms with legal ops teams and IT departments.
You have neither.
You need to know which Claude plan is right for your case mix, which workflows produce real time savings, and exactly where the Florida Bar ethics rules land when you run client information through an AI tool.
This playbook answers all three.
What’s in the playbook
30+ pages, written for firms with no IT department. Here is what you get.
Seven workflow stages, 15 prompts
Intake, research, drafting, review, client communications, matter administration, and marketing. Each stage has specific Claude use cases and hard stops, plus 15 ready-to-adapt prompts calibrated for Florida practice and marked for Claude Chat or Cowork.
The plan decision
Which Claude plans belong on client matters and which do not. Why a personal account is a confidentiality problem, and what to verify in the commercial terms and data processing addendum before client work starts.
Chat, Cowork, and the shift to agents
What Claude Chat does, what Cowork does when it works through a task on its own, and why an agent needs firm-level limits. Includes three immediate opportunities and the governance that goes with them.
Data classification
A four-tier framework that decides which information goes into which plan, from public research to highly sensitive client data, including what changes when a file contains protected health information.
Florida ethics and court rules, applied
The five rules that matter most (4-1.1, 4-1.6, 4-5.3, 4-3.3, and 4-1.5), the other rules that touch AI use, and the current court rules on AI in filings, including Rule 2.515(d)(2) and Administrative Order AOSC26-12. The plain-English version is the 25 AI ethics questions.
Five failure modes and staff training
Consumer-account leaks, hallucinated citations, unsupervised Cowork runs, agent over-reach, and billing inflation, named so you can spot them. Plus the three things everyone in the firm must know before using Claude on a client matter.
Governance policy template
A sample AI governance policy to adapt for your firm: permitted and prohibited uses, confidentiality and data handling, a separate agent policy for Cowork, court filings, incident response, and a staff acknowledgment.
Client and court disclosure language
Engagement-letter language telling clients how the firm uses AI, and court disclosure and certification language matched to the court where you file. Both come with implementation notes and a reminder to have a Florida-licensed attorney review before use.
A vendor letter with nine questions
Nine written questions to send Anthropic or any AI vendor on training, access, storage, encryption, breach notice, and deletion, so the answers go in your governance file instead of staying in someone’s head. For the full vendor review, see the Vendor AI and Data Governance Audit.
A 90-day implementation roadmap
Foundation in days 1 to 30, proficiency in days 31 to 60, optimization in days 61 to 90. It starts from informal individual use and ends with governed, firm-wide use.
The ethics framework
Florida Bar Ethics Opinion 24-1 is explicit: attorneys using AI tools must understand how those tools work and where they fail. That is not optional. It is what competence means now.
Florida Supreme Court Opinion SC2026-0673, effective June 15, 2026, requires every signer of a Florida state court filing to represent that the legal authorities identified exist and are accurately cited. Administrative Order AOSC26-12, effective the same day, bars Florida courts from imposing their own AI disclosure or certification requirements, so the statewide signer representation governs state-court filings. Federal judges still set their own AI orders.
Capital Standard, LLC v. U.S. Bank National Association (Fla. 2d DCA Aug. 21, 2026) made the issue concrete. Counsel’s briefs carried 46 fabricated or unsupported citations. The court struck both briefs, fined him $1,500, made him personally liable for the fees U.S. Bank incurred answering the amended initial brief, and referred him to The Florida Bar. AI-generated citations that do not survive verification are not a harmless tech mistake. They are a lawyer problem. The software will not be standing next to you at the podium. Funny how that works.
- What you can run through Claude
- What requires a verification step
- What requires attorney review before it moves forward
- How to set firm-level rules before staff start improvising
Who this is for
Solo attorneys and small firm partners practicing in Florida. Any practice area. The prompts library covers general firm workflow. No technical background required.
If you are already using AI tools without a governance policy in place, this playbook is overdue.
About Greg Yates
Former Big Law corporate partner. Now solo AI adoption consultant for small law firms.
I do not sell software. I do not get referral fees from AI vendors. I help small firm attorneys build AI workflows that hold up under bar scrutiny.
actionableai.net | [email protected]
Not sure where this tool fits yet? Start with the AI readiness self-check.
Compare the tools
Ready to implement?
Download the full playbook. Or book a 15-minute introductory call if you want to work through your firm’s specific situation: which plan, which workflows, and how to structure governance for your practice area.
No pitch. No deck. No obligation.