Solution · Law firms

AI-first intake with human handoff for legal inquiries

OrcaLinq captures structured intake, pre-screens conflicts, and escalates sensitive matters to a qualified person — without ever giving legal advice.

Why this fits

Intake is repeatable. Legal advice is not.

Law-firm websites lose qualified leads because no one staffs the chat after hours, and most chatbots are too aggressive to be safe in a regulated profession. The semantic PBX handles intake and routing without overstepping into legal advice.

Intake-only AI

AI captures matter type, parties, urgency, jurisdiction, and contact preference. It does not interpret law, recommend action, or imply representation.

Conflict pre-screen

Names captured at intake are flagged for the conflicts team in real time. Final check is human; AI never confirms availability for a matter.

Sensitive escalation

Criminal defence, family-law urgency, restraining orders, or active litigation route immediately to a qualified attorney endpoint — not a junior or a queue.

Audit-grade trail

Wake captures every intake field, every routing decision, every refusal of legal advice. Useful for compliance review, audit, and bar-rule defensibility.

Conversation example

Intake without crossing the line

A typical after-hours inquiry that would otherwise fall on a junior the next morning, often after the prospect has already retained another firm.

Prospect (web widget, Friday 11:42 PM): "I was just served papers for a divorce. What do I do?"

AI: "I'm sorry you're in this. I can't give legal advice, but I can get you in front of a family-law attorney quickly. Could I take a few details to route this correctly? Your full name, your spouse's full name (so we can run a conflict check), the state you were served in, and whether children are involved?"

Prospect: "Sure: [details]"

AI: "Thank you. I'm flagging this as time-sensitive and pinging our family-law team now. Sarah Chen leads that practice and is on call this weekend — she should reach you within the hour. In the meantime, please don't sign anything or post about this on social media. A confirmation email is on the way."

Sarah (WhatsApp from her personal endpoint, 11:47 PM): "Hi [name], Sarah Chen here. I have your intake — let's set a 20-minute call. Earliest is 9 AM tomorrow. Does that work?"

The AI did not advise. It did not promise representation. It captured what a paralegal would, ran the conflict pre-screen, and paged the right partner — with the entire interaction logged in Wake.

Strict limits

What this will never do

  • It will not give legal advice. Period. Refusal is policy-driven, audit-logged, and cannot be jailbroken.
  • It will not commit on confidentiality scope. That requires an engagement letter signed by an attorney.
  • It will not replace conflicts review. AI flags; humans approve. No exceptions.
  • It will not collect privileged content. Intake stays in the public, non-privileged zone until the engagement is set.
  • It will not imply representation. No 'we can take your case' language, ever.
  • It will not predict case outcomes. Even when asked directly. Always escalates.
Recommended setup

What a typical firm wires in week one

  • Website widget on the contact page and every practice-area page, with the practice tag pre-filled.
  • Email channel for prospects who prefer it.
  • WhatsApp for personal-injury, immigration, or family-law firms — where it's culturally expected.
  • Calendar integration for consultation booking.
  • Approved knowledge: practice areas, jurisdictions, hours, fee structure (where public), what to bring to a consult, paralegal staff names.
  • Routing policy: criminal/family/immigration urgency → on-call partner; commercial inquiries → business-development; PR/press → managing partner.
  • Refusal templates for legal-advice attempts, outcome-prediction attempts, and confidentiality-scope questions.
Features used

What's doing the work

Sonar

Reads urgency and matter type. Criminal-defence and time-sensitive matters bypass the standard queue.

Matriarch

Routes by practice area, conflict-check status, and on-call rotation. Out-of-hours and weekend flows are part of policy.

AI human handoff

Forced handoff on advice attempts, sensitive matters, and any topic outside intake. No exceptions, no confidence threshold to override.

Knowledge base AI

Practice-area descriptions, fee structure, and intake fields are the only sources. Nothing answered from general model memory.

Wake

Every intake, every refusal, every escalation logged. Auditable, defensible, exportable.

Metrics to improve

What firms see after launch

  • After-hours inquiries captured — typically zero before, several per week after.
  • Time-to-attorney for sensitive matters — drops from next-morning to under an hour, on weekends and holidays.
  • Front-desk and paralegal time — reclaimed from routine intake calls, redirected to billable work.
  • Conflict pre-screen latency — falls from days to minutes; conflicts team reviews structured records, not voicemails.
  • Lost-prospect rate — measurable when the prospect would otherwise have called the next firm on Google.
Law firms FAQ

Frequently asked questions

Will the AI give legal advice?

No. The platform is configured to refuse legal advice and route those conversations to a qualified person. AI handles intake, scheduling, and conflict-check pre-screening only. The refusal is policy-driven, not prompt-engineered, so it cannot be jailbroken with creative phrasing.

What about confidentiality?

Tenant isolation, encrypted credentials, configurable retention, and a Wake audit trail are all in place. Specific posture — retention windows, redaction rules, on-shore data residency, custom DPA — is configured per firm before launch.

Can it do conflict checks?

It can pre-screen — capture the names and matter type and flag potential conflicts to the intake team in real time. Final conflict checks are always human; AI never tells a prospect 'we can take your case' or implies acceptance of representation.

What channels work for legal?

Website widget and email are the most common; WhatsApp suits firms with personal-injury, immigration, or family-law work where clients prefer mobile messaging. The choice depends on your client base and the formality your firm requires.

Can it schedule consultations?

Yes — when a calendar tool is connected. The AI offers available consultation slots, holds one, and confirms via email and the customer's chosen channel. It will explicitly note that scheduling a consultation is not the same as engaging the firm.

What about after-hours emergencies?

Out-of-hours flows are part of the routing policy. Criminal defence, restraining-order, and time-sensitive immigration matters can page a duty attorney's WhatsApp or phone immediately, not queue until morning.

How does this fit our existing intake process?

It runs in front of it, not instead of it. AI captures structured intake, the same fields your team would on a call, and routes the qualified prospect to the right associate or paralegal with the transcript and intake form pre-filled.

See intake without overstepping.

Bring a typical inquiry your firm receives — including the kind that always tries to extract free advice. We'll show the routing, the intake fields, and the strict refusal of legal advice.