AI legal assistant for lawyers: agents and automation, steps, costs
By Alan Magliocca · Published November 2026
Quick answer: An AI legal assistant for lawyers is a set of automations and AI agents that handle routine steps such as intake summaries, conflict searches, engagement letters, document requests and client status updates. Lawyers must supervise it like staff under ABA Model Rules 5.1 and 5.3, keep a human checkpoint before anything reaches a client or court, and review the vendor's data terms first.
Key takeaways
- Clio reports 72% of solos use AI in some way, but only 8% have built it in widely.
- ABA Opinion 512 applies supervision duties under Rules 5.1 and 5.3 to AI tools.
- Automate hand-offs with plain rules first, then add AI where reading or writing is involved.
- Keep a person between every AI output and a client, court or deadline.
- Read each vendor's data terms before connecting it to client files.
Why automation matters more for solo and small firms
Small practices use AI, but mostly in small ways. In Clio's 2025 report on solo and small firms, 72% of solo legal professionals and 67% of those at small firms said they use AI in some capacity, but only 8% of solos and 4% of small firms had adopted it widely or universally, as summarized in Clio's highlights of the report. These are Clio's survey numbers. The gap between "some use" and "built into how the firm runs" is where automation sits.
For a solo lawyer, the same few chores repeat on every matter: collect intake details, run conflicts, send an engagement letter, open the matter, calendar deadlines, chase documents and answer "any update?" calls. Each is small. Together they eat the hours you meant to bill.
What the ethics rules say about an AI assistant
ABA Formal Opinion 512, issued July 29, 2024, applies the supervision duties in Model Rules 5.1 and 5.3 to generative AI. Lawyers with managerial authority need clear firm policies on approved tools, and lawyers must supervise AI output as they would a staff member's work. The comment to Rule 5.3 already lists using an internet-based service to store client information as an example of outside assistance a lawyer must manage, as the ABA's Rule 5.3 comment explains. In practice, that means reading each vendor's data and security terms before connecting it to client files.
An "AI agent" that can send emails, file documents or move money on its own is a staff member with no judgment. Give it narrow permissions and a person who approves its work.
How to build an AI legal assistant in six steps
- Map one workflow end to end. Start with new-client intake. Write every step from first contact to an opened matter, who does it today and how long it takes. Pick the three slowest steps.
- Choose the system of record. Your practice management software holds matters, contacts and deadlines. Automations should read from it and write to it, not create a second database.
- Automate the hand-offs first. Examples: an intake form creates a contact and a draft conflict search; a cleared conflict creates a draft engagement letter from your template; a signed letter opens the matter and assigns tasks. No AI is needed for these. They are rules.
- Add AI where reading or writing is involved. Let AI summarize a long intake form into five lines for the lawyer, draft a plain-English status update from the matter notes, or sort incoming documents into the right folder. Each output waits for a person to approve it.
- Never fully automate deadlines. Software can suggest dates from court rules, but a lawyer or trained paralegal confirms every calendared deadline against the rule and the court's order.
- Log and audit monthly. Keep a record of what each automation did. Once a month, sample ten actions and check them. Fix the rule, not just the one result.
Calls and website chats are often the front door to this workflow; see AI receptionist and phone answering for lawyers and AI chatbots for lawyers.
Worked example: a solo criminal defense lawyer
This lawyer is fictional. She practices alone in two county courts, and most new clients call within a day or two of an arrest. Her week is broken up by calls from family members asking about the next court date.
She automates in three layers. First, a web intake form feeds her practice software, and an AI summary gives her the charge, the court, the next date and who is paying, in five lines. Second, once she accepts a matter, the engagement letter and fee terms go out for e-signature from her template, and signing opens the matter. Third, each time she enters a court date, a reminder text and email to the client are drafted for her approval, along with a short plain-language note about what that hearing is for.
She does not let AI calculate deadlines or contact anyone without her review. She also confirms with each vendor that client data is not used to train its models.
What an AI legal assistant costs
- By hand: the cost is your time, or a legal assistant's. BLS lists median pay for paralegals and legal assistants at $30.24 an hour in May 2025, before benefits.
- Practice management software: Clio says its plans start at $49 per user per month; workflow automation and e-signature features vary by tier and vendor.
- AI assistant inside your office suite: Microsoft lists Microsoft 365 Copilot Business at about $21 per user per month with annual billing, as an add-on to a Microsoft 365 plan.
- Automation platforms that connect apps: usually priced by the number of tasks run per month. Estimate your monthly volume before choosing a tier.
Mistakes to avoid with legal automation
- Letting an agent send client or court communications without approval.
- Connecting a tool to client files before reading its data terms. Check whether it trains on your data and where it stores it.
- Running without a written AI policy. Clio's 2025 report found 44% of legal professionals said their firm had none.
- Automating a broken process. Fix the steps on paper first.
- Forgetting the exit. Make sure you can export your data if you change vendors.
For lawyers who also sell through events and follow-up, BrightStage AI has a primer on what AI sales automation is.
How AIVX Labs handles automation
AIVX Labs is an AI agency platform, not a practice management system, so it does not hold client files, run conflict checks or calendar court deadlines. Its Automation mapping tool lays out a workflow like the intake example above, step by step, so you can see what to automate. RICO, the platform's AI operator, can run a plan of 2 to 6 tools for marketing and client-growth work with one credit total you approve, and standing jobs repeat that work on a schedule within a credit cap you set. If you use GoHighLevel, results land on the right HighLevel contact with a tag your automations can use. For the money side of the back office, see billing and invoicing for lawyers.
Sources
- Highlights From the 2025 Legal Trends for Solo and Small Law Firms Report (Clio), 2025
- ABA issues first ethics guidance on a lawyer's use of AI tools (Formal Opinion 512), 2024-07-29
- ABA Model Rules of Professional Conduct: Comment on Rule 5.3 Nonlawyer Assistance
- Microsoft Copilot Plans and Pricing for Business (Microsoft)
- Clio Pricing & Plans (Clio)
Related
Written by Alan Magliocca, founder of AIVX Labs and BrightStage AI (Podavinci LLC). Written with AI assistance and checked against the linked sources.
Read our featured article on LinkedIn
Get AIVX Labs verified for $1: RICO, your AI operator, with 7 days of full access.
Get AIVX Labs Verified for $1Follow along:LinkedInYouTubeTikTokXFacebookInstagramSkool community
Frequently asked questions
What is an AI legal assistant?
It is a set of tools that drafts, summarizes and moves information between your systems, such as turning an intake form into a matter summary. It works under a lawyer's supervision, like any assistant.
Can AI agents run my law office on their own?
Not safely. Under ABA Opinion 512 and Rule 5.3, you must supervise the work, so agents should draft and prepare while a person approves anything that reaches a client or court.
Which tasks should a solo lawyer automate first?
Intake, conflict searches, engagement letters and status updates. They repeat on every matter and follow the same steps each time.
Is client data safe in automation tools?
Only as safe as the vendor's terms and settings. Check whether the tool trains on your data, where it stores files and how you can export or delete them.
