Search “AI for law firms” and many top results are lists of AI tools, mostly for research and drafting; we read them in October 2026. Those tools help with the legal work. For a firm owner, the safer first payoff from AI for lawyers is usually the business side: answering new leads, following up, keeping case files complete, telling clients where their case stands, booking consults and reporting what each signed case cost.
This guide maps those six jobs: what AI does, what a person still checks and, where a rule applies, the rule behind that check, quoted from the source. Then it shows how we use AI in our own software and in Command Center, ROI Society’s internal system.
Key Takeaways
- In Clio’s 2025 Legal Trends Report, 79% of legal professionals said they use AI, but 53% said their firm has no AI policy, or they don’t know of one (2Civility summary).
- In Clio’s 2024 secret shopper study, only 40% of law firms answered the phone, down from 56% in 2019 (ABA Journal).
- ABA Model Rule 1.4 requires lawyers to keep clients “reasonably informed about the status of the matter,” and its comment says a lawyer “should promptly respond to or acknowledge client communications” (ABA).
- ABA Formal Opinion 512 says a client’s informed consent is required before information about their matter goes into a self-learning AI tool, and says managerial lawyers “must establish clear policies” on AI use (ABA Formal Opinion 512).
- Our rule: start with one workflow, a written data rule and a vendor check, and keep a person on anything a client, a court or a government agency sees.
On this page:
- The business side first
- Six jobs AI can help run
- What clients expect
- What stays with the lawyer
- How to start
- How we use it
- FAQ
- Sources
AI for law firms pays off first on the business side
AI for law firms covers two kinds of work. One is legal work: research, contract review, document review and drafting legal documents, the tasks ABA Formal Opinion 512 lists. The other is firm operations, the work of running the practice, and that is where this guide spends its time.
In Clio’s 2025 Legal Trends Report, most legal professionals (77%) said improved operations is where they see AI’s biggest revenue gains. Among those using AI, 63% said it improved client responsiveness, according to a summary by the Illinois Supreme Court Commission on Professionalism.
Our take: start on the business side because the output is easy to check. A reminder text or a missing-document list is right or wrong at a glance; a research memo needs a lawyer’s full review. Small firms and solo practitioners gain the most, because the lawyer is often the office manager too.
For drafting and research, see our post on ChatGPT for law firms. For ads, content and review replies, see AI tools for law firm marketing.
Six jobs where AI helps a law firm run today
AI can take the first pass at six jobs that keep a law firm running, and a person still checks the output before it reaches a client. This table is our working summary.
| Job | What AI does | What a person checks |
|---|---|---|
| Intake and answering | Answers calls and chats at any hour, asks a short set of questions, books a consult | Conflicts, urgent calls, anything that sounds like legal advice |
| Follow-up | Drafts missed-call texts, consult reminders and no-show messages | Wording on sensitive matters, texting consent |
| Document and case management | Flags missing or mismatched items, tracks deadlines, drafts replies to incoming notices | Every deadline, and every draft before it goes out |
| Client portal and status updates | Turns the case record into a plain-English update | Any update that gives advice or predicts a case outcome |
| Scheduling | Books, confirms, reminds and reschedules | Court dates and the lawyer’s real availability |
| Reporting | Matches signed cases to their lead source and ad spend | The counting rules behind each number |
Intake and answering: answer fast, ask little, give no advice
AI client intake answers every call or chat, collects the basics and books a consult. It should not tell the caller what will happen in their case.
Take a hypothetical call to a criminal defense firm. A parent calls at 11 p.m. because their son was arrested tonight. AI can take the parent’s name and number, the son’s name, where he is held and any court date, then route the call to the on-call lawyer. It should not guess whether he will get out on bail.
The Florida Bar’s Ethics Opinion 24-1 says a chatbot that talks with clients or others “must include a disclaimer indicating that the chatbot is an AI program and not a lawyer or employee of the law firm.” It also points to older guidance for intake by nonlawyers: ask only for facts, give no legal advice and send legal questions to the lawyer.
The opinion says it is advisory, but we treat it as a sound standard in any state. Our AI receptionist guide covers the call flow and recording consent, and our intake scripts post has the wording.
Follow-up: AI drafts the messages, your team approves them
AI follow-up drafts the text after a missed call, the reminder the day before a consult and the message after a no-show. Your team approves each template once, and the system sends on schedule.
Texts carry consent rules. The FCC says its rules ban autodialed texts to a mobile phone unless the owner “previously gave consent,” and that “Commercial texts require written consent.”
Our rule for family law: the first text names the firm and offers a time to talk, with nothing about the matter, because someone else may read the phone. The cadence itself is in our lead management guide.
Document and case management: AI flags, people decide
In case management, AI is best at catching what is missing or inconsistent: a document not yet uploaded, a name spelled two ways across forms, a deadline coming up. It can also read an incoming notice and draft a reply for staff to review.
How much checking is enough? Opinion 512 says it “will necessarily depend on the GAI tool and the specific task that it performs.” It also suggests marking anything an AI tool produced when you store it in a client or firm file, “so future users understand potential fallibility of the work.”
Client portal and status updates: the duty stays with the lawyer
In our view, status updates are the best fit for AI on the client side, because the rules ask for them. ABA Model Rule 1.4 requires a lawyer to “keep the client reasonably informed about the status of the matter” and to “promptly comply with reasonable requests for information.”
The comment to Rule 1.4 adds that when a prompt answer isn’t feasible, the rule requires the lawyer “or a member of the lawyer’s staff” to acknowledge the request and say when a response may be expected.
A client portal can show the next step and what each side is waiting on; for a criminal defense client, the next court date and what to bring. AI drafts the update from the case record, and a lawyer approves any update that interprets the law or predicts a case outcome.
Scheduling: book, confirm and remind
Our take: scheduling is the simplest job to hand to AI. It books consults on the lawyer’s calendar, confirms them, sends reminders and offers a new time after a no-show. A person keeps court dates and blocked time accurate, because the AI books into whatever the calendar shows.
Measure it with consult show rate: consults held divided by consults booked, one of the KPIs in our law firm dashboard guide.
Reporting: AI matches cases to costs if intake fills the fields
AI reporting can match each signed case to the lead and the ad behind it, then show cost per signed case by campaign. That only works if intake records the lead source and case type every time.
In one law firm account we manage, the county field in the firm’s case system was filled in on 9 of 35 matters in a six-week sample (our data, August 12 to September 24, 2026). So we treat county-level cost per case there as directional and don’t set bids on it.
Required fields beat cleanup. That is why the software we are building won’t let staff mark a case signed until the fee, case type and lead source are in.
Clients now expect fast answers and a view of their case
Would-be clients test a firm’s speed before they meet a lawyer, and many firms fail. In Clio’s 2024 secret shopper study, reported by the ABA Journal, only one-third of 500 firms replied to an email, and only 40% answered the phone, down from 56% in 2019.
Answering was not enough. Of the firms that replied by email, only 18% were clear about next steps and cost. All in all, the article reports, “73% were unlikely to recommend the firms that they contacted.”
Clients are open to AI that speeds up service, and wary of AI that decides. Clio’s 2025 report found, per the 2Civility summary:
- 47% of clients would be comfortable hiring a lawyer who uses AI to respond to clients if it gave the lawyer more time to work on their case.
- 47% would not be comfortable hiring a lawyer who uses AI to make decisions, and 56% would rather their lawyer use a paralegal than AI.
- 78% want lawyers to disclose AI use, yet 35% of legal professionals said they rarely or never disclose it.
- 14% of consumers have used AI to answer legal questions, and 43% haven’t but would. Our GEO strategy guide covers how a firm gets mentioned in those AI answers.
Our take: use AI for speed and visibility, keep decisions with people, and tell clients how you use it.
Judgment, legal advice and supervision stay with the lawyer
Legal judgment, legal advice and supervision stay with the lawyer, even when AI drafts the first pass. Opinion 512 says AI tools “cannot replace the judgment and experience necessary for lawyers to competently advise clients about their legal matters.”
The output also needs checking. Opinion 512 warns that some tools are “prone to ‘hallucinations’” and cites a Stanford study finding that leading legal research companies’ AI systems “hallucinate between 17% and 33% of the time.”
Supervision is where the firm’s owner comes in. Applying Model Rules 5.1 and 5.3, Opinion 512 says “Managerial lawyers must establish clear policies regarding the law firm’s permissible use of GAI.” Supervising lawyers must make reasonable efforts to ensure lawyers and staff meet their professional obligations when they use AI, and that includes training. Our summary: AI can answer, draft and flag; the lawyer decides which cases the firm takes, what gets filed and what a client is told.
Opinion 512 and the Model Rules are the ABA’s guidance; your state bar’s own rules and ethics opinions control, so check them. Our post on AI agents in law firms: security and ethics risks walks through Opinion 512 in full.
Start with one pilot, a written data rule and a vendor check
The safest way to start is small: one workflow, one written rule about client data and one vendor you have checked. Few organizations have a plan. Thomson Reuters surveyed 2,275 professionals in law, tax, accounting, compliance and related fields for its 2025 Future of Professionals report, and only 22% said their organization has a visible, defined AI strategy.
Those organizations were twice as likely as ones with informal or ad hoc adoption to see revenue growth from AI (Thomson Reuters).
These are our six steps:
- List the AI you already use. The Canadian Bar Association says “You are likely already using AI in your legal practice,” because AI now sits inside search engines, browsers and legal databases (CBA). Check your practice management software, phone system and CRM too.
- Pick one workflow with a clear right answer. Consult reminders or status updates make good first pilots. Write down today’s numbers first: answer rate, time to first reply and consult show rate.
- Write the data rule. Name the approved tools and accounts, and what stays out of every other tool. Use placeholders instead of client names when drafting in a general AI tool.
- Get consent where the rule requires it. Opinion 512 says a client’s informed consent is required before you put information about their matter into a self-learning tool, and “merely adding general, boiler-plate provisions to engagement letters” is not enough. Using a tool for ideas, with no client information, needs no consent.
- Check the vendor and train the team. Use the checklist below. Opinion 512 suggests training on “the capabilities and limitations of the tools,” ethics and secure data handling.
- Run it for 30 days, then compare. Name one person who approves every client-facing message during the pilot. Move to the next job only when the numbers and the review notes hold up.
The vendor checklist from Opinion 512
Opinion 512 tells lawyers to read “the Terms of Use, privacy policy, and related contractual terms and policies” of any AI tool they use. It also lists checks from earlier opinions on outside providers and says “These concepts also apply to GAI providers and tools.” The right column is our wording.
| Check from Opinion 512 | What to ask the vendor |
|---|---|
| Terms of use and privacy policy | Who can see what we enter, and do you train on it? |
| Reference checks and vendor credentials | Which law firms use you, and for how long? |
| Security policies and protocols | Where is our data stored, who can access it, and can we see your latest SOC 2 report? |
| Hiring practices | Who on your staff can reach our data, and how are they screened? |
| Confidentiality agreements | Will you sign one that covers our client information? |
| Conflicts check system | How do you screen for adversity among our clients? |
| A legal forum for relief | Where can we enforce the contract if you break it? |
| Configuration and breach notice | Is the tool set up to keep information confidential, and will you tell us about a breach? |
How we use AI in our own software and Command Center
We use AI the way this guide describes: it drafts, checks and reminds, and a person approves anything a client, a court, a government agency or an ad account sees.
Who does what. Pallazola Law, a new firm starting with immigration law, is owned and run by attorney Wen Pallazola; ROI Society handles marketing, growth and business operations. Wen makes every legal decision. ROI Society has no ownership stake, takes no share of legal fees and does not practice law.
The case management software. We are building it, and Pallazola Law is the first firm on it. As of October 2026 it is in development, and no real client has used it yet. “Built” below means working in our test setup (full feature list).
| Feature | What it does | Status, October 2026 |
|---|---|---|
| Intake | Branching questionnaire; no case marked signed without the fee, case type and lead source | Built |
| File checks | 14 automatic consistency checks; staff accept each upload or send it back | Built |
| Deadlines | Deadline calculators and a reminder to check each case’s status every 7 days | Built |
| Client portal | Next task, questionnaire, uploads, messages and calendar, in English and Simplified Chinese | Built; client sign-in and messages not on yet |
| Notice reader | Reads USCIS notices and requests for evidence and drafts replies for staff to approve | Built, switched off |
| Staff helper | Answers staff questions from the staff manual; gets no client records and strips personal details | On in our test setup, with a daily spending cap |
AI that reads client documents stays off until the firm accepts the AI provider’s data-retention terms. Every database table has access rules, and documents sit in private storage.
Command Center. Command Center is ROI Society’s internal system of record: clients, tasks, support tickets, documents and the history of every ad account change. A daily check proposes ad changes, and none goes live until a person approves it. By Jay Pallazola’s own estimate (October 2026), reports that took him about 100 hours now take 5 to 10 hours.
Each client firm’s portal shows only its own leads, signed cases and cost per signed case; our law firm dashboard post shows how it fits together. For client firms, we set up AI chat, voice and follow-up in GoHighLevel through our custom AI automations and CRM and marketing automation work.
If you can’t see where new leads stall between the first call and a signed case, a Law Firm Growth Evaluation walks through it with you.
Frequently asked questions about AI for law firms
Can we charge clients for our AI tools?
In our reading of Opinion 512, usually not for tools that run the office. A tool that works like “equipping and maintaining a legal practice” is overhead unless you told the client otherwise in advance. Lawyers who bill hourly “must bill for their actual time.” Opinion 512 covers flat fees too, which matters for criminal and family firms: if AI lets you finish “much more quickly,” charging the same flat fee may be unreasonable under Rule 1.5. Our take: when AI cuts the work on a flat-fee matter type, review that fee.
Do we have to tell clients we use AI?
It depends on the facts. Opinion 512 says “The facts of each case will determine whether Model Rule 1.4 requires lawyers to disclose their GAI practices to clients.” Our take: tell clients in plain words what AI does in your office and what a person checks.
Are there AI lawyers?
Not in the sense of a licensed lawyer. Florida’s bar opinion says a law firm’s chatbot must state that it is “an AI program and not a lawyer or employee of the law firm.” Software sold as an AI lawyer can draft and explain, but a licensed lawyer gives the advice and answers for it.
Is ChatGPT safe for client information?
Not by default. Opinion 512 tells lawyers to read a tool’s terms of use and privacy policy “to learn who has access to the information that the lawyer inputs into the tool,” and matter information in a self-learning tool needs the client’s informed consent. Otherwise, use placeholders instead of names. Our ChatGPT for law firms post covers settings and safer uses.
Will AI replace our intake staff or paralegals?
Our take: no. AI takes the repetitive first pass, so your team spends more time with clients and less retyping. Someone still has to clear conflicts, calm a scared caller and decide what goes to the lawyer.
Use AI to run the firm and keep the law with the lawyer
AI for law firms works best where the work repeats and is easy to check: intake, follow-up, files, status updates, scheduling and reporting. Start with one job, write the data rule, check the vendor and keep a person on what clients see. The legal judgment stays where the rules put it, with the lawyer.
Sources
- American Bar Association, “Formal Opinion 512: Generative Artificial Intelligence Tools” (July 29, 2024), https://www.americanbar.org/content/dam/aba/administrative/professional_responsibility/ethics-opinions/aba-formal-opinion-512.pdf, checked October 21, 2026.
- American Bar Association, “Rule 1.4: Communications,” https://www.americanbar.org/groups/professional_responsibility/publications/model_rules_of_professional_conduct/rule_1_4_communications/, checked October 21, 2026.
- American Bar Association, “Rule 1.4 Communication, Comment,” https://www.americanbar.org/groups/professional_responsibility/publications/model_rules_of_professional_conduct/rule_1_4_communications/comment_on_rule_1_4/, checked October 21, 2026.
- Illinois Supreme Court Commission on Professionalism (2Civility), “2025 Clio Legal Trends Report Outlines Firms’ Tech and AI Use” (October 29, 2025), https://www.2civility.org/2025-clio-legal-trends-report/, checked October 21, 2026.
- ABA Journal, “With lawyers being less responsive to clients, Clio unveils new AI-productivity aid to help” (October 7, 2024), https://abajournal.com/news/article/with-lawyers-being-less-responsive-to-clients-clio-unveils-new-ai-productivity-aide-to-help, checked October 21, 2026.
- Thomson Reuters, “The AI Adoption Reality Check: Firms with AI Strategies are Twice as Likely to see AI-driven Revenue Growth; Those Without Risk Falling Behind” (June 26, 2025), https://www.thomsonreuters.com/en/press-releases/2025/june/the-ai-adoption-reality-check-firms-with-ai-strategies-are-twice-as-likely-to-see-ai-driven-revenue-growth-those-without-risk-falling-behind, checked October 21, 2026.
- The Florida Bar, “Ethics Opinion 24-1” (January 19, 2024), https://www.floridabar.org/etopinions/opinion-24-1/, checked October 21, 2026.
- Federal Communications Commission, “Stop Unwanted Robocalls and Texts,” https://www.fcc.gov/consumers/guides/stop-unwanted-robocalls-and-texts, checked October 21, 2026.
- Canadian Bar Association, “2. Use of AI in Legal Practice: An Overview,” Ethics of Artificial Intelligence for the Legal Practitioner, https://cba.org/resources/practice-tools/ethics-of-artificial-intelligence-for-the-legal-practitioner/2-use-of-ai-in-legal-practice-an-overview/, checked October 21, 2026.
Header image: AI-generated image.
See what turns your marketing
into signed cases.
Your firm doesn't need more random marketing. It needs a growth system that shows what turns into signed cases.