Two Lawyers, One Paralegal, $94 a Month: Inside the AI Makeover of a Small Law Firm
Most small law firms don't have a software problem. They have a software pile. That was certainly the case at Henderson & Ruiz, a two-partner family law and estate planning practice in a suburb of Columbus, Ohio, that let us audit every tool, subscription, and sticky-note workaround in their office.
The numbers weren't pretty: $1,140 a month on software for a five-person firm, much of it for features nobody could locate on a screenshot. Over six weeks, we rebuilt the stack around AI tools that actually pull their weight. Final tab: $94/month. Here's how we got there — and where we almost got it wrong.
The Before Picture: $1,140 a Month of Bloat
The legacy stack looked like this:
- Practice management suite (the big-name kind): $499/month
- Legal research subscription: $389/month
- Document assembly add-on: $99/month
- Transcription service: $85/month
- Legacy billing software: $68/month
The practice management suite was the worst offender. Partners Elena and Marcus used maybe 15% of it. Intake forms lived in a separate tool. Billing was exported to Excel anyway. And the research subscription? Marcus admitted he "just Googles most things and panics."
The Rules We Set
Before touching anything, we established three ground rules — and if you run a firm, write these down:
- No client data in consumer AI tools. Ever. This shaped every choice below.
- Anything billing-adjacent must keep an audit trail. No black boxes.
- The paralegal, Dani, has veto power. She's the one who'd live in these tools daily.
That third rule mattered more than the tech. Our first draft got vetoed twice.
The New Stack: Tool by Tool
Document Review & Drafting: Claude Team — $30/user/month
We tested Claude, ChatGPT Team, and Gemini for Business against real firm work: summarizing 80-page custody filings, flagging inconsistencies in financial disclosures, and first-drafting demand letters.
Claude won on long-document accuracy and tone. The Team plan includes a training opt-out, which matters enormously here — more on that below. Two seats: $60/month.
Transcription: Otter Pro — $17/month
Deposition prep meetings and client calls get transcribed automatically. Otter isn't perfect with legal jargon, but Dani edits as she goes, and the time savings beat the $85 legacy service handily. $17/month.
Intake & Scheduling: Jotform + Cal.com — $34/month
The old suite's intake module was replaced with a Jotform feeding straight into Cal.com scheduling. Intake-to-appointment time dropped from two days to about four hours. $34/month total.
Billing: Kept the Legacy Tool — $0 saved (for now)
Surprise: we tried to replace billing with a modern AI-assisted invoicing tool, and Dani's veto kicked in. Trust matters more than features when money touches an IOLTA-adjacent workflow. We kept the existing billing software and cut it from the comparison. Sometimes the honest answer is "don't switch."
Research: Google + a State Bar Database — $0
The $389 research subscription covered mostly federal case law the firm almost never touches. Ohio-specific resources through the bar association plus careful Google Scholar use covered 90% of actual needs. Marcus now uses Claude to summarize cases he's already found — not to find them. That distinction matters.
Running total: $94/month (Otter annual pricing) plus the retained billing tool.
The Confidentiality Trap That Almost Sank Everything
Here's the part every firm owner needs to hear. In week two, Dani started pasting redacted client agreements into a free AI chatbot to speed up review. Perfectly reasonable instinct. Potentially catastrophic decision.
Free and even some paid consumer AI plans may use your inputs for training. For an attorney, that's not just a privacy issue — it's a potential ethics problem under ABA Model Rule 1.6 on client confidentiality, and several state bars have issued specific guidance on generative AI.
The fix: every AI tool in the stack is on a business/team tier with a contractual training opt-out, and we wrote a one-page firm AI policy. Redaction before anything touches a tool, even a compliant one. If you take one thing from this article, take this paragraph.
The Before/After Budget, Line by Line
| Line Item | Before | After |
|---|---|---|
| Practice management | $499 | $0 (replaced) |
| Legal research | $389 | $0 (bar resources) |
| Document assembly | $99 | — (folded into Claude) |
| Transcription | $85 | $17 (Otter) |
| Billing | $68 | $68 (kept) |
| Claude Team (2 seats) | — | $60 |
| Jotform + Cal.com | — | $34 |
| AI intake tool trial | — | $0 (canceled) |
| Total | $1,140 | $179 |
Wait — $179? Right. The $94 figure counts only the new AI stack (Claude, Otter, Jotform/Cal.com at annual rates). Add back the retained billing tool and you get the true monthly total. We're being picky about this because every vendor "savings" story on the internet rounds in its own favor.
What We'd Tell Your Firm
- Cut the suite last, not first. Migrate intake and docs before you cancel anything.
- Business tiers only. The training opt-out isn't optional in legal work.
- Budget for human editing. AI drafts, humans sign. Nothing goes out the door unreviewed.
- Keep billing boring. Boring billing software is a feature, not a failure.
Six weeks in, Elena and Marcus report saving roughly 10 hours a week across the firm — and the $961 in monthly savings is funding a part-time contract attorney. That's a better ROI than any software feature we could name.
