Frequently Asked Questions

LawFirm AI ships with pre-built intake flows for the twelve most common plaintiff-side practice areas:

• Personal injury (auto, slip-and-fall, premises liability, product liability)
• Mass torts and class actions
• Family law (divorce, custody, domestic violence)
• Immigration (removal defense, visa petitions, asylum)
• Criminal defense (felony, misdemeanor, DUI)
• Employment law (wrongful termination, discrimination, wage theft)
• Workers' compensation
• Social Security disability (SSDI/SSI)
• Estate planning and probate
• Bankruptcy (Chapter 7 and 13)
• Real estate disputes
• Civil rights litigation

Defense-side and transactional practices can use our custom flow builder to define their own qualification criteria. Enterprise plans include white-glove flow authoring by our legal operations team.

Escalation is governed by a four-trigger system you configure during onboarding:

• Score threshold: When a prospect's qualification score (0-100) crosses the threshold you set, the AI alerts your on-call intake coordinator in real time via SMS and email.
• Distress signals: The AI is trained to detect urgency markers -- statute of limitations within 30 days, active court dates, incarceration, safety risk language -- and immediately flags these regardless of score.
• Explicit request: If the prospect says "I want to talk to someone" at any point, the AI warm-transfers to your intake line or schedules a callback at the prospect's preferred time.
• Confidence floor: If the AI encounters a question or scenario outside its training, it says so and routes to a human rather than guessing. We call this the "honest hand-off" policy.

All escalations are logged with full conversation context so the receiving attorney or coordinator never has to ask a prospect to repeat themselves.

Yes. LawFirm AI detects the prospect's language automatically from the first message and responds in kind. We support 47 languages at launch, with Spanish, Portuguese, Mandarin, Vietnamese, Korean, Tagalog, and French receiving the deepest legal-domain fine-tuning.

For intake flows in languages you have attorney capacity for, the qualified-lead packet is delivered in both the prospect's language and English. For languages where you cannot serve the client, the AI politely explains this and offers a referral pathway -- keeping your firm compliant with professional responsibility rules around language-based discrimination.

It understands enough to qualify -- and knows exactly where its jurisdiction ends. The model was fine-tuned on over 2.3 million anonymized intake transcripts across 14 practice areas, which means it recognizes the difference between a garden-variety fender-bender and a catastrophic injury case, or between a simple uncontested divorce and a high-conflict custody situation with domestic violence history.

What it will not do: render legal opinions, advise on strategy, predict outcomes, or discuss pending case strategy. It is an intake specialist, not a lawyer. Every user-facing interaction includes a clear disclosure that the AI is not an attorney and the conversation does not create an attorney-client relationship. This language is reviewed by our legal advisory board and kept current with ABA Formal Opinion 512 guidance.

At the start of every intake session, the AI collects opposing-party information before the prospective client discloses any confidential facts. This pre-screen is cross-referenced against your existing client database via our Clio, MyCase, or Filevine integration in real time.

If a potential conflict is detected, the session is suspended immediately -- before any material facts are shared -- and the prospect is informed that your firm cannot assist them at this time. The flagged session is quarantined in a separate log visible only to the conflict-clearance partner. Your main intake team never sees the substantive conversation. This architecture is designed to satisfy the screening walls required under ABA Model Rule 1.10.

They receive an immediate, empathetic response. The AI opens by acknowledging the situation, collecting the critical facts, and assessing urgency. If the case is time-sensitive -- an active arrest, a custody emergency, a statute of limitations within 48 hours -- your designated emergency on-call contact receives a priority alert within 90 seconds.

For non-urgent after-hours leads, the AI completes the full intake, scores the case, and presents the prospect with two options: schedule a callback during business hours (calendar invite sent instantly) or receive a preliminary case summary by email so they feel heard and informed before the call. Either path converts at significantly higher rates than a voicemail greeting or a generic "we will call you back" message.

Fully. The scoring engine exposes every dimension as a configurable weight. Common customizations include:

• Minimum injury severity threshold (e.g., ignore soft-tissue-only cases below a certain treatment cost)
• Geographic restriction (e.g., only qualify cases in states where you are licensed)
• Liability threshold (e.g., require at least one clear adverse party before scoring as qualified)
• Case-value floor (e.g., auto cases with damages under $10,000 score below the alert threshold)
• Urgency multipliers (SOL within 60 days doubles the priority score)

Changes to scoring weights take effect in under two minutes and do not require a support ticket. Enterprise clients also get access to our A/B testing framework, which lets you run two scoring configurations simultaneously and compare conversion rates before committing to a new model.

No. Your clients' intake conversations are never used to train, fine-tune, or evaluate any model -- ours or a third-party vendor's. This is a hard contractual commitment in our Data Processing Agreement, not a policy that can be toggled by an engineering team. The training data for our models consists entirely of consented and anonymized historical intake records licensed directly from law firms under IRB-equivalent review.

All inference happens on isolated, single-tenant compute for Enterprise plans. For Growth plans, the inference layer is multi-tenant but the data storage layer is fully isolated per firm with separate encryption keys.

Every connection to and from the LawFirm AI platform uses TLS 1.3 with certificate pinning for mobile clients. Data at rest is encrypted with AES-256 using keys managed through AWS KMS. Enterprise clients can bring their own key (BYOK) and rotate on their own schedule.

Intake conversation content is stored separately from identifying metadata, with cross-referencing only available to authenticated, role-permissioned users within your firm. Database-level encryption with field-level encryption for the most sensitive fields (SSN, DOB, contact info) is enforced by default on all tiers.

Current certifications and compliance posture:

• SOC 2 Type II Annual audit by a Big Four firm, report available under NDA
• HIPAA BAA available for medical-legal and mass tort practices handling PHI
• CCPA Data subject rights portal live for all California residents
• GDPR Data Processing Agreement available; EU data residency option on Enterprise
• Pen Test Annual third-party penetration test, most recent report dated Q1 2026

We publish a live Trust Center at trust.lawfirmai.com where you can view current uptime, incident history, and the latest audit summaries without waiting for a sales call.

AI intake tools are permissible in all 50 U.S. states and D.C. under the unauthorized practice of law rules, provided the AI does not render legal opinions and the supervising attorney maintains oversight of the intake process. LawFirm AI is architected specifically to satisfy these requirements: it collects facts, qualifies cases, and schedules consultations -- all activities courts and bar associations have consistently held to be administrative, not legal, in nature.

We track ethics opinions in all 50 states and update our product and disclosure language accordingly. Our legal advisory board includes former state bar ethics committee chairs from New York, California, Texas, and Florida. An ethics compliance summary report for your state is available upon request and is included in all Enterprise onboarding packages.

The AI makes four disclosures at the start of every session, presented in plain language before any information is collected:

• That the user is speaking with an AI, not a human attorney or paralegal
• That the conversation does not create an attorney-client relationship
• That the firm's privacy policy governs how the information will be used
• That the user may request to speak with a human at any time

Disclosure language is customizable by your firm's ethics counsel and is displayed as both a chat message and a persistent banner throughout the session. Acknowledgment of the disclosures is logged with a timestamp as part of the intake record.

We ship native (no-code) two-way integrations with:

• Clio Manage -- matters, contacts, tasks, and documents sync automatically
• MyCase -- leads, contacts, and intake questionnaires
• Filevine -- projects, contacts, and custom fields
• Smokeball -- matters and contacts
• Lawmatics -- full CRM pipeline, custom stages, and automation triggers
• Law Ruler -- leads, pipeline status, and call scheduling
• Salesforce -- via the Salesforce for Legal AppExchange package

For systems not on this list, we offer a documented REST API and a Zapier integration that covers 5,000+ downstream apps. Enterprise clients receive a dedicated integration engineer for the first 90 days at no added cost.

Yes. Deployment is a two-line JavaScript snippet for the web widget -- works with WordPress, Squarespace, Webflow, custom React apps, and any other web platform. The chat widget is fully white-labeled with your firm's name, colors, and logo.

For phone, we integrate with Twilio, RingCentral, and Dialpad. Prospects who call your intake line during off-hours can be routed to an SMS-based intake flow, a voice-based AI intake assistant (available on Enterprise), or a scheduled callback system. All three options are configured in your dashboard without engineering support.

Yes. Every new firm gets a 30-day free trial on the Growth plan, including up to 250 live intake sessions and two CRM integrations. No credit card is required to start. At the end of the trial, you receive a full performance report showing leads qualified, escalations triggered, and estimated case value captured -- so you can make an ROI-grounded decision before spending a dollar.

We do not shut off your intake. Every plan includes burst capacity that allows you to exceed your session limit by up to 30% without interruption. Overages are billed at a flat per-session rate ($1.20 on Growth, $0.80 on Scale, $0.55 on Enterprise) at the end of the billing cycle. You receive an email alert at 80% and 95% of your limit so there are no surprises.

If you consistently exceed your plan's sessions for two months running, your account manager will proactively reach out about an upgrade -- not a penalty.

Month-to-month plans cancel with 30 days notice, no questions asked, no cancellation fee. Annual plans can be cancelled with a 60-day notice if the firm changes size, merges, or closes -- in those circumstances we waive the remaining term. We do not charge cancellation fees on annual plans for any other reason.

Upon cancellation, your firm retains a full data export of all intake records in JSON and CSV format. Data is retained in our systems for 90 days post-cancellation, then permanently deleted per our retention policy.

Most Growth-plan firms are live within one business day. The onboarding flow is self-serve and takes approximately 45 minutes:

• Choose your practice areas and connect your CRM (15 min)
• Review and customize the pre-built intake flows (15 min)
• Configure escalation thresholds and notification contacts (10 min)
• Paste the widget snippet on your website and test a live session (5 min)

Scale and Enterprise plans include a dedicated onboarding engineer who handles CRM configuration, custom flow authoring, and staff training. These onboardings typically complete within five business days.

No engineering background required. The dashboard was designed to be operated by intake coordinators, paralegals, and office managers -- not IT departments. All configuration is done through visual interfaces with plain-language labels. The only step that requires developer access is the initial website snippet installation, which takes under two minutes for any web developer or site admin to complete.

Based on aggregate data from 340 firms that completed their first 90 days on the platform:

• 38% average increase in qualified leads reaching an attorney consultation
• $4.20 recovered per $1 spent in the first 90 days, driven primarily by after-hours lead capture
• 27% reduction in intake coordinator time spent per qualified lead
• 11-day reduction in average time from first contact to retained client

Results vary by practice area and volume. Personal injury and mass tort firms see the steepest initial gains due to high after-hours lead volume. Family law and criminal defense typically see the largest time-savings gains. Your account manager will build a custom ROI projection during onboarding based on your current intake data.

The platform tracks 22 quality metrics across four categories. Your weekly digest email surfaces the eight most actionable:

• Completion rate -- percent of sessions that reach a scored outcome vs. drop off mid-intake
• Qualification accuracy -- percent of AI-qualified leads that attorneys confirm as genuinely qualified after review (our benchmark: 91%)
• False positive rate -- percent of AI-qualified leads that attorneys reject as unqualified
• Escalation timeliness -- median time from urgency trigger to human response
• Prospect satisfaction score -- post-session survey (1-5 scale) sent to 100% of completed intakes
• Conversion rate -- percent of qualified leads that become retained clients
• After-hours capture rate -- percent of total leads arriving outside business hours that were successfully qualified
• Conflict detection rate -- percent of sessions flagged for potential conflict before material facts were disclosed