What a law firm AI receptionist actually does
Answers an intake call, screens the caller by practice area and conflict-check details, and books a consultation or routes the call to an attorney, without ever giving legal advice.
A law firm AI receptionist runs an intake call the way a trained legal intake specialist would: greet the caller, find out what kind of matter they have, capture the facts needed for a conflict check (names of opposing parties, case type), and either book a consultation on the calendar or route the call to an attorney. The one hard boundary that separates this from every other receptionist on this site is legal advice: the agent can gather facts and schedule a meeting, but it can never answer a legal question, estimate an outcome, or tell a caller what to do, because doing so risks the unauthorized practice of law.
A 2024 Clio Legal Trends Report secret-shopper study of 500 firms found that only 40% answered the phone and just 33% responded to email, with 73% of the callers who did get through saying they wouldn't recommend the firm they reached. That gap is exactly what Smith.ai's native Clio integration and Upfirst's Clio calendar sync are built to close, and it's why legal intake has become one of the more mature verticals for AI receptionists even though the overall market for this exact keyword is small.
Where it still breaks: anything that edges into giving legal advice, no matter how reasonable the caller's question sounds, and conflict checking itself, which the agent can capture data for but can never clear on its own. The ABA's Formal Opinion 512 treats an AI vendor as a nonlawyer assistant under Rules 5.1 and 5.3, meaning the firm supervising it is responsible for its output exactly as if it had hired a person to answer the phone.
AI capabilities involved
Real-time speech-to-text on the live call
Screening by practice area and urgency
Capturing conflict-check facts without clearing the conflict itself
Booking a consultation via tool calls into the firm's calendar
Who uses this
- Solo and small-firm attorneys (1-10 lawyers) losing potential clients to voicemail or a missed call
- Personal injury and family-law firms with high call volume and a defined intake script
- Firms already running Clio Manage or Clio Grow wanting phone intake that writes into the same system
- Firms whose consultation calendar changes constantly and needs real-time availability, not a callback queue
What the law firm receptionist handles (and what it shouldn't)
| Task | Agent handles it? | Systems it touches | Human role |
|---|---|---|---|
| Answering every call so nothing goes to voicemail | Fully | Phone number, telephony carrier | None |
| Screening by practice area and urgency | Fully | Intake script | Reviews unclear or borderline cases |
| Capturing conflict-check data (names, case type) | Fully | Conflict-check tool, CRM | Runs the actual conflict check |
| Booking a consultation | Fully | Calendar (Clio, Calendly) | None for standard slots |
| Answering "do I have a case" or any legal question | No | None | An attorney answers directly |
| Existing-client case-status updates | With approval | Case management system | Confirms before sharing any details |
| Payment reminders and billing questions | With approval | LawPay, billing system | Reviews disputed charges |
| Angry or distressed callers | No | Call transfer | An attorney or office manager takes over immediately |
Off-the-shelf AI agents on the market
Real products you can sign up for today — with current 2026 pricing, honest pros and cons.
Smith.ai AI Receptionist
Small law firms already on Clio wanting the deepest legal-specific integration
Free plan: 25 calls/mo, $3/call overage
Pro $150/mo (75-300 calls, about $2/call); Enterprise $500/mo (300-1,000+ calls, about $1.67/call)
Enterprise tier scales past 1,000 calls/mo
Pros
- +Native Clio integration built specifically for law firm intake
- +Month-to-month with 30-day notice, no annual lock-in
- +Native Zapier and Calendly connections for firms not yet on Clio
Cons
- −Per-call pricing costs more at volume than a flat-rate alternative
- −The free tier's $3/call overage adds up fast past 25 calls
Upfirst
Firms wanting the lowest published starting price
14-day trial
Starter $24.95/mo (30 calls); Premium $59.95/mo (90); Pro $159.95/mo (300); Scale $299/mo (600)
Not published; contact sales for higher volume
Pros
- +Cheapest published entry price of any vendor on this page
- +Syncs with Clio, Google, Outlook, Acuity, Calendly and Square calendars
- +Supports 35+ languages out of the box
Cons
- −Overage runs $1.50 down to $0.70/call, which nearly doubles the per-call cost at the Starter tier
- −No case-management-specific tools beyond calendar sync
LawDroid (Copilot)
Firms wanting to pilot AI intake over chat before adding a voice line
Not published
Copilot reported at roughly $25/user/mo; its intake and builder tiers are reported inconsistently across sources at $49-$299/mo
Not published
Pros
- +Purpose-built for legal work, including document drafting alongside intake
- +A chat-first option if you want to test AI intake before committing to the voice channel
- +Legal-industry focus means less generic scripting than a horizontal receptionist tool
Cons
- −Pricing figures for its intake and builder products conflict across third-party sources
- −Chat-first positioning means it's less proven specifically as a phone receptionist
LEX Reception (human receptionists, for comparison)
Firms wanting a human safety net for complex intake rather than AI
None
Unverified third-party figures: roughly $425/150 min, $450/300 min, $775/500 min
Not published
Pros
- +Bilingual human receptionists with native Clio and MyCase integrations
- +A useful cost benchmark for comparing AI pricing against a live-human answering service
- +No AI-disclosure concerns since it's genuinely staffed by people
Cons
- −It's a human service, not AI, so it's the expensive baseline this category exists to beat
- −Pricing isn't confirmed on the vendor's own public materials
The AI stack
A law firm receptionist needs the same voice and telephony layers any phone agent does, plus a calendar and conflict-check integration and a hard-coded legal-advice boundary no model should be trusted to enforce on its own.
Voice (speech-to-text + text-to-speech)
Converts the caller's speech to text and the agent's reply back to speech
ElevenLabs Agents
$0.08/min on every self-serve plan; $0.16/min burst over the concurrency capCustom builds prioritizing voice quality
Deepgram + OpenAI Realtime
Deepgram STT plus gpt-realtime-2.1 at roughly $0.04-$0.10/min combinedDeveloper-assembled stacks
Our pick: ElevenLabs plus a chained reasoning model for most custom builds, since legal intake doesn't need the lowest-latency speech-to-speech models.
Telephony
Carries the call and connects it to your voice pipeline
Lindy's built-in phone agents
Included in Lindy creditsNo-code builds on Lindy
Twilio direct
$0.008/min inbound, $0.014/min outboundFully custom builds
Our pick: Let your chosen platform handle telephony pass-through rather than managing Twilio directly, unless you already run Twilio elsewhere.
Reasoning model
Screens the caller, holds the conversation, and enforces the no-legal-advice boundary
Claude Sonnet 5.5
$2/M input, $10/M outputThe main intake conversation loop
Claude Haiku 5.5
$0.10/M input tokens (up to 100K), $0.50/M outputRoutine confirmation calls
Our pick: Sonnet 5.5 for the main intake loop, with a separate rule layer, not the model's judgment alone, blocking any answer that strays into legal advice.
Calendar & conflict-check integration
Books a real consultation slot and captures the facts a human will use to run the actual conflict check
Clio Grow intake and calendar
Included in a Clio Grow subscriptionFirms already on Clio
Google Calendar or Calendly plus a CRM field
Free to low-costFirms not yet on a legal-specific CRM
Our pick: Use Clio Grow's intake tools if you're already on Clio; otherwise capture the same fields into whatever CRM your firm already uses.
Observability & evals
Reviews call transcripts for any answer that drifted toward legal advice
Vendor call logs and transcripts
Included on most SaaS tiersA starting point, not a substitute for attorney review
Langfuse plus attorney spot-checks
Free up to 50K units/mo, then $29+/moAny production legal deployment
Our pick: Have an attorney, not just an office manager, review a weekly sample of calls specifically for anything that edges toward legal advice.
Reference architecture
The hardest engineering problem is the legal-advice boundary itself: a caller will often ask a question phrased as a logistics question ("how long will this take", "what are my chances") that's actually a request for legal judgment, so the architecture needs a hard rule, not the model's own confidence score, deciding when a question has to go to an attorney.
Call connects
TelephonyThe caller's phone call routes to the receptionist's number
Speech-to-text
STT engineThe caller's audio is transcribed to text in real time as they speak
Practice-area and advice-boundary check
Reasoning model plus a hard-coded rule layerThe model identifies the practice area and intake facts; any question that edges toward legal advice trips a hard rule regardless of model confidence
Conflict-check data capture
CRM/conflict-check toolNames, case type and other conflict-relevant facts are logged for a human to actually clear
Calendar tool call
Clio Grow or calendar integrationA consultation slot is booked directly against real attorney availability
Transfer or escalation
Call-transfer logicLegal-advice-adjacent questions, distressed callers, or anything outside the script get transferred to an attorney
Call logging and eval
Observability toolEvery call is logged, and an attorney reviews a weekly sample specifically for advice-boundary drift
Estimated cost per request
~$0.08-$0.15 per minute of call on a self-assembled stack; roughly $1.67-$2.00 per call on packaged SaaS tiers
Cost calculator
Drag the sliders to model your actual usage. The numbers update in real time so you can stress-test economics before writing a single line of code.
Estimate your monthly cost at your own call volume and average call length.
Estimated monthly cost
$253
≈ $3,036 per year
Calculator notes
- The per-call line uses Smith.ai's roughly $2/call Pro-tier rate as a conservative proxy across packaged vendors.
- The DIY line assumes a 3-minute average call on an all-in Lindy or Vapi-style voice stack at roughly $0.11/min.
- A custom agency build replaces both lines with your own voice-stack bill, typically $50-$300/mo at this volume, plus the upfront build cost.
- None of these figures include attorney time spent on the actual conflict check or consultation itself.
Build it yourself with vibe-coding tools
Lindy lets a solo practitioner assemble a working intake line with phone agents and calendar integrations, without writing a backend or paying a per-call vendor fee.
Time to MVP
A few days for a basic version, 2-4 weeks to get the legal-advice boundary and conflict-check capture production-ready
Total cost to MVP
$0-$50 (Lindy's Team plan at roughly $30/mo for 3,000 credits, plus a phone number)
You'll need
Starter prompt
Create a new phone agent in Lindy using Claude Sonnet 5.5 as the model. Connect a calendar tool (Clio Grow, Google Calendar or Calendly) and a CRM or sheet for conflict-check data. Set the first message to: 'Thanks for calling {Firm Name}, this is their AI assistant, how can I help you today?' Give the agent this system prompt: 'You are the intake receptionist for {Firm Name}, a law firm handling {practice areas}. Greet callers, find out what kind of legal matter brought them to call, and collect the names of all parties involved, the general case type, and their contact information for a conflict check, using the conflict-check tool to log it. Use the booking tool to schedule a consultation with an available attorney. You must never answer a legal question, never estimate how a case will turn out, never give an opinion on whether someone has a case, and never suggest a course of action, even if directly asked; if a caller asks any of these, say: "That's something one of our attorneys needs to answer directly, let me get you scheduled," and proceed to booking. If the caller sounds distressed or the matter sounds urgent, use the transfer_call tool to reach the office directly instead of booking a later slot. Never claim to be human if asked directly. This call is recorded for quality purposes, consistent with the greeting already given.' Add three tools: log_conflict_check, book_consultation, and transfer_call pointed at your office line. Turn on call recording, and have an attorney review a weekly sample of transcripts specifically for any answer that strayed toward legal advice.
Paste this into Lindy
Follow-up prompts (run in order)
- 1
Add a Spanish-language path that detects the caller's language and switches voices
- 2
Add a tool that checks whether a potential client's statute of limitations is close to expiring and flags it for same-day attorney review
- 3
Add an SMS confirmation sent to the caller after every booked consultation
- 4
Add a nightly report listing every call transferred and why
Expected output
A phone number that screens callers by practice area, captures conflict-check data, books a real consultation slot, and transfers anything that edges toward legal advice directly to an attorney.
Known gotchas
- !The legal-advice boundary has to be a hard rule, not something left to the model's judgment, or it will eventually answer a question it shouldn't
- !ABA Opinion 512 treats the AI vendor as a nonlawyer assistant under Rules 5.1 and 5.3, meaning the firm is responsible for its output exactly as if it had hired a person
- !Lindy's credit-based billing is harder to forecast against call volume than a flat per-minute rate
- !Forgetting the AI-disclosure line at call start is both a compliance problem and a common caller complaint
- !A conflict check still has to be run by a human; the agent only captures the data a person needs to run it
- !Call transfers need a warm handoff with context, or the attorney picking up has to ask the caller to repeat everything
Compliance & risk reality check
A legal intake agent carries a different kind of compliance weight than other receptionists: the risk isn't primarily a regulator, it's the firm's own professional-responsibility rules.
ABA Formal Opinion 512 (supervision and confidentiality)
Issued July 29, 2024, Opinion 512 treats an AI vendor as a nonlawyer assistant under Rules 5.1 and 5.3, requires informed client consent before putting confidential information into a self-learning tool, and bars billing for time saved by using it.
Mitigation: Have a supervising attorney review the agent's scripts and a sample of its transcripts regularly, use a vendor with a no-training clause, and disclose AI use in engagement letters.
Unauthorized practice of law
The agent's job is intake, not advice; answering a legal question, estimating an outcome, or recommending a course of action risks the unauthorized practice of law regardless of how the caller phrased the question.
Mitigation: Hard-code a transfer trigger on any legal-advice-adjacent question, independent of the model's own confidence.
Conflict-check data capture, not clearance
The agent can and should capture the names and facts a conflict check needs, but it can never clear a conflict or decide the firm can take a case.
Mitigation: Route every captured conflict-check record to a human for the actual clearance before the firm engages a new client.
Call recording and wiretap law
Ambriz v. Google and Taylor v. ConverseNow both survived motions to dismiss in 2025 on the theory that an AI vendor capable of using call data to train its own models can be an unlawful eavesdropper.
Mitigation: Disclose recording and AI at the start of every call, and bar every vendor in the chain from training on your call data by contract.
TCPA for outbound follow-up calls
If the firm has the agent call back a missed intake lead, that's an outbound AI voice call under the FCC's February 2024 ruling (24-17), which requires prior consent and disclosure.
Mitigation: Capture written consent before any outbound AI callback, and disclose the AI at the start of the call.
Build vs buy: the real math
3-6 weeks for a RapidDev build; a few days for a basic Lindy version
Custom build time
$13K-$25K for an agency build (up to $35K for multi-system or compliance-heavy work); $0-$50/mo for DIY
One-time investment
Roughly 400-500 calls a month, where per-call SaaS fees start to exceed the amortized first-year cost of a custom build
Breakeven vs buying
At 100 calls a month, Smith.ai's roughly $2/call Pro rate runs about $200/mo, well under a $13K-$25K custom build's amortized year-one cost of $1,083-$2,083/mo. The math shifts for firms with heavy call volume (a high-intake personal injury practice, for example) or firms running several distinct practice areas with different intake scripts and conflict-check needs, where a custom build's flat model cost beats per-call pricing. For a typical solo or small firm, buying stays the right call well past where this page's cost calculator starts to feel uncomfortable.
Skip the DIY — RapidDev builds the production version
A Lovable MVP gets you a demo. Production needs auth that doesn't leak data, AI calls that don't bankrupt you, observability when models drift, and code you can audit. That's what we ship through our AI app development service.
Discovery call (free)
30 minWe map your exact law firm AI receptionist use case: who uses it, target volume, AI model choice, integrations, compliance scope. You get a detailed scope document and fixed-price quote within 48 hours.
AI-accelerated build
3-6 weeks for a RapidDev build; a few days for a basic Lindy versionOur engineers use Claude Code, Lovable, and custom tooling to ship 3–5x faster than agencies. You see weekly progress in a staging environment — not a black box.
Launch + handoff
1 weekWe deploy to your infrastructure, transfer the GitHub repo, set up CI/CD and monitoring, and train your team. You own 100% of the source code, prompts, and model configurations.
What you get
Timeline
3-6 weeks for a RapidDev build; a few days for a basic Lindy version
Investment
$13K-$25K for an agency build (up to $35K for multi-system or compliance-heavy work); $0-$50/mo for DIY
vs SaaS
ROI in Roughly 400-500 calls a month, where per-call SaaS fees start to exceed the amortized first-year cost of a custom build
30-min call. Fixed-price quote within 48 hours. No commitment.
Frequently asked questions
How much does an AI receptionist for a law firm cost?
Packaged tools run $24.95-$500/mo depending on call volume: Smith.ai is $0-$500/mo by call, Upfirst is $24.95-$299/mo by call. A custom RapidDev build runs $13K-$25K upfront (up to $35K for complex, multi-system work), plus $50-$300/mo in ongoing voice-stack costs. A DIY build on Lindy costs roughly $30/mo plus model and phone costs.
How long does it take to set up?
A packaged tool like Smith.ai or Upfirst can go live the same day you connect your Clio account. A custom agency build takes 3-6 weeks to wire into your practice areas and conflict-check workflow. A basic Lindy version can work within a few days, though getting the legal-advice boundary production-ready takes 2-4 weeks.
Can RapidDev build this for us?
Yes. RapidDev has shipped 600+ apps and offers a free 30-minute consultation to scope your practice areas, call volume and compliance needs before you commit to anything.
Can the AI receptionist give legal advice?
No, and it should never try. The agent's job is screening, conflict-check data capture and scheduling; any question that edges toward legal advice, an outcome estimate or a recommended course of action has to transfer to an attorney immediately, regardless of how confident the model sounds.
Does ABA Opinion 512 apply to an AI receptionist?
Yes. Opinion 512 treats an AI vendor as a nonlawyer assistant under Rules 5.1 and 5.3, which means the supervising attorney is responsible for what the agent says and does, exactly as if a person had been hired to answer the phone.
Can it work with Clio?
Yes. Smith.ai has a native Clio integration built for legal intake, and Upfirst syncs directly with Clio's calendar alongside Google, Outlook, Acuity and Calendly. A custom build can connect to Clio Grow or Clio Manage directly through their APIs.
Will callers know they're talking to AI?
They should hear it in the first line of the greeting. Stating "this is their AI assistant" satisfies disclosure expectations under California's BOT Act, Utah's AI Policy Act and the EU's Article 50 transparency rule at once, on top of being good practice for a first-time caller to a law firm.
Want the production version?
- Delivered in 3-6 weeks for a RapidDev build; a few days for a basic Lindy version
- You own 100% of the code
- AI cost monitoring built in
30-min call. No commitment.
