AI Client Intake for Law Firms: Capture, Qualify, and Conflicts-Check 24/7

Most managing partners can describe the perfect intake — a structured interview, a same-day conflicts check, an engagement letter signed before the prospect calls a competitor. The intake actually running inside the firm rarely looks like that. A web form fills at 8:47 p.m., sits in an unmonitored inbox, gets glanced at the next morning, routed to whichever associate has bandwidth on Wednesday, and the prospect — who already filled out three other firms' forms — has signed with a competitor by Thursday afternoon.

That is the lead-response problem, and it is what AI intake-automation was built to fix. Every credible study of legal lead conversion since the original Harvard Business Review/InsideSales response-time research lands on the same point: prospects who get a substantive response in under five minutes convert at 21x the rate of prospects who get a response in over thirty. Most SMB firms respond in 4–28 hours.

This piece is the operational walkthrough — vendors, call flow, conflicts logic, bar-compliance posture, 9-day rollout. The strategic frame sits in the AI for lawyers playbook and the dollar math in the ROI walkthrough.

The lead-response problem in legal

The math is not subtle. Clio's Legal Trends Report and Lawmatics benchmarks converge on a few numbers.

• 35–48% of legal leads arrive outside 9-to-5 hours; a third land after 8 p.m.
• The median SMB firm response time to a web form is 12.8 hours. The median to an after-hours call is "voicemail, returned next business day."
• Firms responding under 5 minutes convert to consultations at 28–42%. Over an hour, the rate drops to 2–7%.
• 67–74% of prospects who fill out a contact form fill out at least one competitor's form. The first credible response wins.

For most firms the binding constraint on revenue growth is not marketing spend, not website conversion, not close rate — it is lead-response-time. AI intake automation is the only lever that moves it without hiring a 24/7 staffed desk.

The AI intake flow, end to end

The reference architecture for SMB-firm AI intake has three channels and one orchestrator.

Channel 1 — Conversational web form

When a prospect lands on the contact page, the form is a conversation, not a 14-field PDF replacement. The agent (Lawmatics, Clio Grow, or a custom build on Claude) asks for matter type, opposing party, jurisdiction, and timeline in a chat-style flow that takes 90 seconds. Drop-off is 40–60% lower than a traditional form because the prospect feels heard rather than processed.

Channel 2 — Voice agent for after-hours and overflow

Phone calls outside business hours and overflow during business hours route to an ai-receptionist. Smith.ai, RUBY (now Posh), Numa, and custom voice agents on Vapi or Retell handle the SMB tier. The agent answers in under two rings, runs the same structured intake the chat agent runs, and books a consultation directly into the partner's calendar via the practice-management API. Calls are transcribed and saved to the matter file.

Channel 3 — SMS intake from paid-ad landing pages

For firms running Google LSA, paid search, or paid social, SMS intake outperforms phone and web. Lawmatics and Clio Grow both support SMS-first intake flows. The prospect texts a keyword; the AI agent runs the interview by SMS; the prospect ends with a scheduled consult.

The orchestrator

All three channels write to one matter record. Lawmatics is the dominant orchestration layer for SMB firms because conflicts logic, engagement-letter templates, e-signature, and CRM live in one place. Clio Grow plays the same role for firms standardized on Clio Manage. The orchestrator runs four downstream actions automatically.

• Soft conflicts check. Cross-references the prospect, opposing party, opposing counsel, and named affiliates against the matter database and the conflicts list. Soft hits get flagged for the conflicts attorney. The engagement letter never sends on a soft hit.
• Practice-area routing. Matter type plus jurisdiction maps to the right attorney. A New Jersey employment matter never lands with the Delaware corporate partner.
• Engagement-letter draft. Pulled from the firm's library, parametrized with matter-specific scope, fee structure, and conflict-waiver language. Queued for partner review and one-click send.
• Consultation booking. Calendar block on the right attorney with a Zoom link, reminder cadence, and pre-consult intake summary.

Named vendors and where they fit

The 2026 SMB legal intake vendor stack has stabilized.

• Lawmatics. Best-in-class legal CRM and intake automation. Conflicts, engagement letters, e-signature, and nurture flows in one platform. Pricing $199–$499 per user per month. The default recommendation for most 2-to-15 attorney firms.
• Clio Grow. Tight integration with Clio Manage. Slightly thinner on the marketing-automation side than Lawmatics but the cleanest data handoff to Clio Manage. $109–$179 per user per month.
• Smith.ai. Hybrid AI-plus-human receptionist. AI handles structured intake; humans handle escalations and emotionally sensitive calls (family law, criminal). Per-call pricing $7–$11.
• RUBY (now Posh). Premium human-led receptionist with AI augmentation. Better fit for firms that explicitly want a human first voice on the line.
• Numa. Voice-first AI receptionist; strong on call recording, transcription, and after-hours coverage. Pairs well with firms running on PracticePanther or Centerbase where Lawmatics integration is thinner.
• CallRail. Call tracking and conversation intelligence. Not an intake vendor, but the analytics layer that proves the intake stack is working — and the source of truth for marketing-attribution disputes.

A typical 5-attorney firm runs Lawmatics or Clio Grow as the orchestrator, Smith.ai or Numa for voice, and CallRail underneath for measurement. All-in monthly cost: $1,400–$2,900.

The 9-day rollout

Same cadence as the broader AI playbook, focused on intake.

• Days 1–2 — Audit. Pull 90 days of every inbound lead. Bucket by channel (web, phone, SMS, referral). Calculate current lead-response-time by channel and by hour-of-day. Compute current set-rate (lead-to-consultation-scheduled) and show-rate.
• Days 3–4 — Configuration. Stand up Lawmatics or Clio Grow. Load matter-type questions, conflicts list, engagement-letter templates, and routing rules. Connect the calendar and PMS via API. Configure Smith.ai or Numa call flows.
• Days 5–6 — Shadow mode. Live traffic runs through the AI; outputs go to the intake coordinator for human review before they reach the prospect. Tune routing, tone, and the conflicts logic.
• Day 7 — Test calls. Five staff-initiated test calls and five test web-form submissions across matter types. Verify conflicts hits, engagement-letter accuracy, calendar routing.
• Day 8 — Cut-over. AI handles all inbound. Intake coordinator monitors a queue rather than answering.
• Day 9 — Measure. Twenty-four-hour response time, set-rate, and show-rate vs baseline. Most firms see response time drop from hours to under 60 seconds and set-rate lift 8–15 points.

Pitfalls and bar-compliance posture

Four risks are worth taking seriously.

Unauthorized practice of law (UPL)

The AI agent cannot give legal advice. It cannot tell a prospect whether they have a case, estimate damages, or recommend a course of action. The flow is information capture and routing only. Every AI script gets a closing line equivalent to "an attorney will follow up with you to discuss your matter" and stops there. State bars are converging on the view that AI intake is permissible when scoped to information collection and scheduling.

Conflicts before engagement

The engagement letter cannot send before the soft conflicts check clears. Make this a hard system gate — Lawmatics and Clio Grow both support it. A firm that automates engagement-letter sending without the gate is a Rule 1.7 violation waiting to happen.

Confidentiality of intake content

Intake data is privileged information from the moment the prospect engages, even if no engagement letter is ever signed. The vendor must sign a confidentiality agreement and ideally a Business Associate-equivalent. Consumer-tier voice agents that train on call audio fail Rule 1.6. Lawmatics, Clio Grow, Smith.ai, and Numa all have appropriate confidentiality language; consumer products do not.

Disclosure to the prospect

Increasingly, state bars expect prospects to know they are interacting with an AI agent for the intake. The disclosure is a one-line script at the start: "Hi, this is the firm's intake assistant — I'll get the basics and route you to the right attorney." That is sufficient under current guidance.

Metrics that matter

Five numbers determine whether the intake stack is working. Track all five monthly.

• Response time, p50 and p95. The headline metric. Target p95 under 2 minutes across channels.
• Set rate. Leads that schedule a consultation. Target lift 8–15 points within 60 days.
• Show rate. Scheduled consults that actually attend. Target 70%+ with AI reminder cadence.
• Engagement rate. Consults that sign. Should not move much from AI intake — the consult and close are still attorney work — but it can drift if the AI is over-qualifying or under-qualifying. Watch it.
• Cost per signed matter. Total intake spend divided by signed matters. Target reduction 20–30% as set-rate climbs faster than spend.

FAQ

Q: Will clients hate talking to an AI? A: 2024–2026 data is consistent: prospects prefer a fast AI response to a slow human one. When the alternative is voicemail, AI wins on satisfaction. The disclosure script at the start handles transparency.

Q: What about emotionally sensitive matters — family law, criminal defense? A: Run AI for triage only. Route sensitive matters to a human within minutes rather than letting the AI conduct a long interview. Smith.ai's hybrid model is built for this.

Q: How does the AI handle a clear conflict? A: Soft hit flags the matter for the conflicts attorney before any engagement letter sends. The AI never tells the prospect "we cannot represent you" — that is attorney communication.

Q: Can I run intake automation without changing my PMS? A: Yes. Lawmatics, Clio Grow, and Smith.ai integrate with major PMSs via API. You do not need to migrate Clio Manage to deploy Lawmatics on top of it.

Q: What about leads that should not be signed? A: The AI can run a fit-screen at the end — minimum matter value, jurisdiction, statute-of-limitations window — and route unfit prospects to a referral list. Protects partner calendar time more than any other single tactic.

Q: How does this compare to a full-time intake coordinator? A: A fully loaded coordinator runs $58k–$78k. The AI stack runs $17k–$35k annually with no PTO gap. Most firms run both: AI for first response, human for warm handoffs.

Q: What is the realistic conversion lift? A: 8–15 set-rate points within 60 days is the median across SMB firms we have measured. At $40k average matter value, that is $400k–$750k of incremental revenue for a $25k stack.

---

If you want a sized intake build against your actual lead volume, matter mix, and PMS — reach out and we will walk it line by line. Or start with the AI for lawyers engagement overview.

SOURCES

Cited and consulted.

1. 01Lawmatics Blog — Legal CRM and Intake Benchmarkslawmatics.com · accessed May 8, 2026
2. 02Clio Legal Trends Reportclio.com · accessed May 8, 2026
3. 03Law.com — Legal AI and Intake Coveragelaw.com · accessed May 8, 2026
4. 04Above the Law — Legal Innovation Centerabovethelaw.com · accessed May 8, 2026