24/7 Voice AI Intake for a Law Firm.

An eight-attorney litigation firm was doing what almost every firm of that size does. Calls came in during business hours to a shared front desk, and after hours they went to voicemail.

The numbers behind that arrangement are unkind. A Law Leaders study of roughly 1,200 calls found that around 35 percent of calls to small and mid-size firms go unanswered even during business hours. Separate research on legal buying behaviour puts 78 percent of clients hiring the first firm that responds to them. And around 85 percent of people who reach a voicemail at a service business never call a second time.

For this firm the arithmetic ran roughly like this. About 300 inbound calls a month. A third of them at risk. Around half of those from genuine prospective clients rather than vendors, opposing counsel or wrong numbers. At an industry retention rate near 7 percent and an average retained case value around $8,000, the missed half of the phone line was the single largest line item nobody was tracking.

There was a second cost. Attorneys were fielding advertising and screening calls themselves, roughly an hour a week each, at a $250 hourly rate.

Delivered in collaboration with Visionary Automate, a systems-integration partner of Zealous Digital Solutions.

A voice agent answers 100 percent of inbound calls, at any hour, on the first ring. That is the whole premise, and everything else is downstream of it.

The agent handles advertising and spam calls in both directions. Inbound solicitation is screened and ended without reaching an attorney. Outbound verification calls, where the firm needed to confirm a detail before a caller was routed, run through the same layer.

Qualification is written against the firm's own case criteria rather than a generic intake script. The agent establishes matter type, jurisdiction, timing and conflict-relevant details, and it declines politely and clearly when a matter falls outside what the firm takes. That refusal path mattered as much as the acceptance path, because a firm that says yes to everything on the phone spends attorney hours saying no later.

Everything the agent captures syncs into the CRM automatically, so a qualified caller arrives as a record with a written summary rather than as a sticky note. No attorney re-keys anything.

The qualification logic and the CRM integration were specified and built jointly. Delivered in collaboration with Visionary Automate, a systems-integration partner of Zealous Digital Solutions.

The figures here are modeled on the firm's stated volumes and published industry benchmarks. They are not audited results. The assumptions:

• Roughly 300 inbound calls per month • 35 percent at risk of going unanswered, per Law Leaders call-study benchmark data • About half of at-risk calls from genuine prospective clients, or roughly 52 a month • 85 percent of unanswered callers never calling back, per the same benchmark set • A 7 percent retention rate on qualified prospective callers • An average retained case value of $8,000 • Attorney time valued at $250 per hour, roughly 1 hour a week each across eight attorneys

Those inputs model out to roughly 3.6 additional retained cases a month, or about $346,000 a year, plus roughly $100,000 in recovered billable capacity from screening time. Total modeled annual benefit exceeds $400,000, and the model suggests payback inside weeks. Actual results depend on the client's baseline and adoption. These figures are modeled estimates, not measured client results.

The front desk stopped being a bottleneck. Two calls arriving at once is no longer a decision about which one to lose, and a lunch hour is no longer a gap in coverage.

Intake quality became consistent, which was the unplanned benefit. A human intake conversation varies with who takes it and how busy the morning is. The agent asks the same qualifying questions every time, which means the CRM now holds comparable records, and the firm can see for the first time what kinds of matters are actually calling.

After-hours calls stopped being a category. An injury or an arrest does not wait for nine in the morning, and those callers previously reached voicemail at exactly the moment they were most likely to hire whoever answered.

Attorneys stopped screening. That hour a week per attorney was not just cost, it was interruption in the middle of billable work, which is the more expensive half of the loss.

The escalation rules that protect those hours were written into the handover documentation as a joint deliverable. Delivered in collaboration with Visionary Automate, a systems-integration partner of Zealous Digital Solutions.

This fits a law firm of roughly 3 to 25 attorneys handling consumer-facing matters where speed of response decides who gets retained: personal injury, family, criminal defence, immigration and employment. Call volume of 150 a month and up is where the model starts to hold together.

It is aimed at United States firm owners and managing partners, and it is region agnostic, because the constraint is the phone rather than the geography. The qualification logic has to be written against the firm's real case criteria, so a firm without written criteria will need to define them first.

It is a poor fit for firms doing purely referral or institutional work, where inbound calls are rare and every caller is already known.

Four weeks, in a fixed order, and the firm receives a written artifact at the end of each one.

• Week 1, discovery and criteria. We sit with whoever currently answers the phone and with an attorney who decides which matters the firm takes. Deliverable: a written intake specification listing every matter type the firm accepts, every one it declines, the jurisdictions it covers, and the exact questions that separate the two. • Week 2, build. The agent is written against that specification, connected to the CRM, and given the firm's own decline language rather than a generic refusal. Deliverable: full transcripts of test calls covering an accepted matter, a declined matter, a conflict-flag matter and a solicitation call. • Week 3, supervised pilot. The agent takes overflow and after-hours calls while the front desk keeps daytime coverage. Every transcript is reviewed by a named person at the firm. Deliverable: a correction log and a revised intake specification, because week 3 exists to find the questions the specification missed. • Week 4, cutover. The agent takes 100 percent of inbound calls, CRM sync runs unattended, and escalation paths for urgent matters go live. Deliverable: a call-handling runbook and the list of conditions that route a caller to a human immediately.

Nothing goes live in week 1. A voice agent talking to prospective clients before an attorney has read its transcripts is a risk no firm should accept.

Five prerequisites. The first two are the ones firms most often discover they do not have.

• Written case criteria. The agent qualifies against the firm's real acceptance rules, so those rules have to exist on paper before they can exist in software. A firm that has never written them down will spend week 1 doing that, and it is worth the week. • A phone system that can forward or port a number. Most can. Some older on-premise systems cannot forward selectively by time of day, which changes the design. • A CRM or case management system with an API, so a qualified caller lands as a record rather than as a task somebody transcribes. • A named person who owns escalation. When the agent flags a caller as urgent or as a possible conflict, somebody has to be reachable and accountable within a stated window. • A decided position on call recording and consent for every jurisdiction the firm takes calls from. This is a legal question the firm answers, not a technical one we answer for them, and it has to be settled before any recording is enabled.

What happens when the agent cannot handle a caller?

It stops trying and routes the call, and it says so plainly rather than improvising. Anything touching legal advice, fee quotes, an existing matter or a distressed caller is out of scope by design, and those calls transfer or generate an immediate escalation with the transcript attached. The refusal path was specified before the acceptance path, because a firm is harmed far more by an agent that answers a question it should not than by one that hands off too often.

Who owns the phone number and the call data?

The firm does. The number stays in the firm's account, transcripts and recordings sit in storage the firm controls, and the CRM is the firm's existing system. If the agent is switched off, calls route back to the front desk and every record captured stays where it already is.

What drives the ongoing running cost?

Call volume first, then how long the average call runs, then the retention period for transcripts and recordings. A firm handling 300 calls a month with two-minute qualifications sits well below one handling the same volume with long intake conversations. Integration count matters less here than in a data build, because most firms connect one CRM and nothing else.

How is success measured in the first 90 days?

Four numbers, three of which the firm can capture before launch. Percentage of calls answered, number of qualified prospective clients reaching the CRM per month, attorney hours spent screening, and time from first call to first attorney contact. Retained cases are the number everyone wants, but it lags the intake cycle, so it is read at month six rather than month three.

Some firms should not do this, and the reasons are worth stating plainly.

• Firms doing purely referral or institutional work, where inbound calls are rare and every caller is already known by name. • Firms with no written case criteria and no appetite to write them. Qualification against undefined rules produces confident sorting of the wrong things. • Practices where the first contact must legally or ethically come from a licensed attorney. In those settings the agent should be scoped to scheduling and message capture only. • Firms that have not resolved recording and consent obligations across the jurisdictions they take calls from.

Naming these is not modesty. A firm that buys this with any of them unresolved gets a build that underperforms and a compliance question it did not plan for.

Delivered in collaboration with Visionary Automate, a systems-integration partner of Zealous Digital Solutions.

The firm is not named and no jurisdiction, practice location or client matter is identified. The build described here is real and in production.

The revenue figures are not measured results. They are modeled from the firm's stated call volume, retention rate and average case value, combined with published call-study benchmarks from Law Leaders where noted. Actual results depend on the client's baseline and adoption. These figures are modeled estimates, not measured client results.

If your firm is sending calls to voicemail after five, or losing the second caller whenever two arrive together, that is the exact gap this build closes. Start a conversation with your monthly call volume, your practice areas and your average retained case value, and we will model it against your numbers rather than these.