Most intake AI improvises. This one cannot. Every answer traces back to a line an attorney signed off on, anything outside that goes to a human mid-conversation, and the whole thing leaves a record you can hand to your carrier or your bar.
Illustrative transcript. The blocked handoff is real product behavior, not a script.
The same exchange, at the same moment, from both sides. Everything on the right is produced automatically by what happens on the left — the caller is never asked to fill in a form, and never sees a score, an entry number, or a flag.
No scores. No entry numbers. No mention of a queue.
Every line above is on the audit trail, with a name and a timestamp.
Both halves of that are true, and most intake tools only solve the first one.
Injury calls come at night and on weekends, when the caller is hurt, frightened, and working down a list of firms. Voicemail is a lost case.
Ask most intake bots what a case is worth and they will give an answer. That answer becomes an expectation the firm has to live with, or walk back.
Bar guidance increasingly treats an AI assistant as something a lawyer has to supervise. That is difficult when there is no record of what it was permitted to say.
The assistant gets better every week, and it gets better only when a named person decides it should.
Every response is retrieved from answers your firm wrote and approved. There is no general knowledge fallback. If it is not in the library, the assistant does not know it.
Below your confidence threshold, it hands off. On a blocked topic it hands off instantly regardless of confidence. Everything the caller already said is saved on the lead, so nobody starts over.
Anything it could not answer becomes an item waiting for review, with a draft answer already written. Approving one takes about fifteen seconds on a phone.
An approved answer goes live immediately and carries who approved it and when, permanently. That record is the difference between a tool you use and a tool you can defend.
Most of the anxiety about putting AI on an intake line comes from not knowing where its edges are. So the edges are a setting, not a promise.
Hours, location, parking, case types handled, what to bring, how contingency fees work in plain language. Answered instantly, every time, from your approved wording.
Answered only above the threshold you set. Start it conservative and watch the handoff rate in your dashboard. Loosen it when you trust it, not before.
Case value. Filing deadlines. Whether someone has a case. Fault. What to tell an adjuster. Checked before anything else, and confidence is not consulted. It hands off. Always.
The moment a caller gives their name, a conflict check is raised on it — matched against every prior intake and every party already on the record, and put in front of a human either way. What happens next is a person's decision, recorded with their name on it.
Software that decides a matter is clean is software making a call that belongs to a lawyer. A missed conflict is a disqualification, a malpractice claim, and a bar complaint. That is not a feature worth automating away, so this does not.
Similar name on a closed 2023 matter. Different date of birth on file.
There is no status in this system that a machine can set to cleared.
The approved answer library. Add, edit, retire. Every entry carries its version number and the name of who approved it.
Your blocked topic list, with the reason recorded next to each one.
A confidence threshold you move, with the resulting handoff rate shown next to it.
Greeting, tone, and the exact words it uses when handing off to a person.
By topic, by hour, by confidence, or on request. The caller can always just ask for a person.
Staff can draft. Only an attorney can approve an answer or resolve a conflict.
Stated as what is true today. What is not yet in place is listed underneath, in the same size type.
Each firm runs on its own database, on its own install. Your intake data does not sit in a pool shared with other firms, and no query in the application can reach from one firm's records into another's.
There is no generative model anywhere in the answering path, so no caller content is sent to a model provider and none of it can become training data. That is a property of how this is built, not a promise about someone's retention policy.
Per-user accounts, no shared logins, and a role split enforced in code: staff can draft an answer, only an attorney can approve one or resolve a conflict. Failed sign-ins are rate limited and lock out.
Every approval, block, retirement, conflict resolution, and settings change is written to an append-only audit log. The application has no code path that edits or deletes a row in it, and a removed blocked topic is deactivated rather than destroyed.
Transport is TLS in front of the application. Encryption at rest, a formal backup and retention policy, and a signed data processing agreement are committed work and are not in place today. If any of those is a gating requirement for your firm, ask where it stands before you sign anything — including here.
A completed intake is a conversation where a caller gave a name and a way to reach them. Bots, wrong numbers, and one-word hangups are not billable and never have been. Overage is 2 dollars per intake, printed here rather than buried in an order form.
Setup and knowledge base build: $1,495, one time. That is the work of sitting down with your team and writing the answers the assistant is allowed to give.
No, and it is built so it cannot. Advice-shaped questions are on the blocked list, checked before anything else runs, and they produce a handoff rather than an answer.
It says so and brings in a person. It does not guess, and it does not fill the gap with something plausible. The question then shows up in your queue.
Any time, in any wording. That path is never gated, delayed, or discouraged.
The honest answer is that it can only say what your firm approved, which is why the approval record matters. The full allocation belongs in the agreement, in writing, reviewed by your carrier.
Chat is live today. Voice runs on the same approved library, the same blocked topics, and the same approval queue, and lands in the same lead record. Do not sell voice until it ships.
Per topic or per answer from the dashboard, without calling support. Retire an answer and the assistant stops saying it on the next message. Your approved answer library belongs to you.
The demo runs against a real approved library. Ask it something it knows, then ask it what your case is worth and watch what it refuses to do.