The AI intake synopsis: never triage a lead blind again

Here's how a lead usually reaches an attorney: a web form or a voicemail becomes an email, the email gets forwarded, somebody re-keys the details into whatever system the firm uses, and eventually a lawyer opens the file cold. If the prospect attached documents, they're sitting unread in the thread. The attorney's first fifteen minutes with the lead are spent doing triage a machine could have done — figuring out what kind of case this is, whether it's urgent, what's missing, and whether the firm can even take it.

Multiply that by every inquiry, including the ones that were never going to become clients, and intake becomes one of the most expensive unbilled activities in the firm. The fix isn't a prettier form. It's making the system do the reading before the attorney spends a minute on the file.

Structured from the first click

EdgeLex intake starts with a form builder — field types, conditional logic, validation, and file upload — published as public hosted forms. The prospect tells their story into a structure, not into a free-text email, and their documents arrive attached to the inquiry rather than scattered across a thread. Nothing gets re-keyed, because it was never unstructured to begin with.

Every inquiry lands in a CRM-style pipeline — new, contacted, qualified, converted — with source attribution, assignment, tags, and contact history. That's table stakes, and it matters mostly for what it enables: once the lead is structured data in your system, the AI can go to work on it.

A qualification score you can interrogate

Each lead is scored and classified: practice area, urgency, and a confidence level — with the reasoning attached. That last clause is the whole point. A black-box lead score is just someone else's opinion wearing a number. A score with its reasoning attached is something a lawyer can actually use, because you can do to it what you'd do to an associate's recommendation: ask why.

If the score says high urgency and the reasoning says the prospect described events suggesting a limitations problem, you know what to check first. If the confidence is low, you know the classification is a starting point, not a verdict. High-urgency leads surface for review instead of waiting their turn in the queue — so the inquiry that can't wait doesn't sit behind the ones that can.

The synopsis: the documents get read, not just stored

This is the piece that changes the attorney's first minutes with a lead. When a prospect uploads documents, EdgeLex reads them — and builds a living synopsis of the lead:

• A summary of the working picture — what this inquiry actually is.
• Key facts extracted from the story and the documents: parties, relationships, the shape of the dispute.
• Risk signals worth an attorney's attention.
• Missing facts — what the file doesn't yet establish, so the first conversation asks the right questions.
• Recommended next actions.
• The potential matter or matters this lead should become.

Consider a hypothetical from the Hale v. Northstar Logistics universe: a freight broker submits an inquiry about a contract dispute with a carrier, and attaches the master agreement and a string of demand letters. Before anyone at the firm opens the file, the synopsis has the parties, the contract's key positions, the timeline the letters establish, a flag that the agreement contains a fee-shifting clause worth reading early, and a note that the submission never says whether the carrier has already filed suit. The attorney walks into the first call knowing the file — and knowing exactly which question to ask first.

Two details in that scenario deserve emphasis. First, the missing-facts list is doing as much work as the extracted facts: the most expensive intake mistake is the question nobody asked until week three, and a synopsis that says what the file doesn't establish turns the first consultation into a targeted interview instead of a fishing expedition. Second, the synopsis is a living picture, not a one-time report — as the lead progresses and more arrives, the working picture is there for whoever touches the file next, so the second attorney to look at the lead doesn't start from zero either.

The AI reads the intake documents so the attorney doesn't triage blind — key facts, risks, and gaps before the first phone call.

Conflicts checked while the prospect is still a prospect

Every lawyer knows the bad version of this story: the conflict surfaces after the consultation, sometimes after the engagement letter, and now the firm is unwinding a relationship it never should have started. The timing is the failure — conflicts checked late are conflicts discovered late.

EdgeLex runs conflict-party detection against your clients, adverse parties, and related entities before engagement — while the prospect is still a prospect. And it doesn't pretend conflict analysis is fully automatable: judgment calls route to a human review queue. The machine does the sweep; the lawyer makes the call. What the firm gets is the guarantee that the sweep happened, on every lead, before anyone was engaged — and the supervisory dashboard flags any lead whose conflict check is missing.

Return to the hypothetical: the freight broker's dispute turns out to be with Northstar Logistics — the same Northstar the firm is already litigating against in Hale. The sweep catches the name against the firm's adverse parties before anyone schedules a consultation, and the relationship question lands in the review queue for a lawyer to evaluate on the facts. Maybe it's a real conflict, maybe it isn't — but the decision gets made by a lawyer, deliberately, at the only point in the timeline where every answer is still cheap.

From qualified lead to open matter, without re-keying

A qualified lead moves onto a single rail. Schedule the consultation, and a prep task comes attached — the synopsis means preparation is reading a briefing, not assembling one. Then the retainer runs through its own desk: requested, document prepared, signed, collected into trust, with reminders when it stalls. The gap between "we'll send the retainer" and money in trust is where engagements quietly die; a desk that tracks each stage and nags the stalled ones closes it.

Conversion is one click. The lead becomes a matter with fields pre-populated from intake and the submitted documents carried over. Nothing is re-typed, no attachment is lost in a forwarded thread, and the lineage runs unbroken from the first form submission to the open file. The documents the prospect uploaded on day one are in the matter's DMS on day thirty — because they never left the system.

Supervision built in

AI in the intake pipeline raises a fair question: who's watching it? The supervisory dashboard answers the questions a managing partner would actually ask. Where is the pipeline bottlenecked? Which leads are stuck or stale? Which are missing conflict checks? Which retainers are unsigned? Which high-urgency leads has nobody reviewed?

Those last two are the ones that should keep a partner honest. An AI that flags a lead as high urgency has done its job only if a human looks — and the dashboard makes an unreviewed urgent lead visible instead of quietly aging in a queue. Meanwhile, every stage change, contact, and conversion is logged on the lead's activity trail, so the pipeline isn't just supervised in aggregate; each lead carries its own record.

What this adds up to

None of these pieces is exotic on its own. Structured forms exist. Lead pipelines exist. What changes the economics is the sequence: the story and documents arrive structured, the AI reads and briefs before an attorney touches the file, conflicts run before engagement rather than after, and conversion carries everything forward without a single re-keyed field. The attorney's time moves from triage — figuring out what this is — to judgment: deciding what to do about it.

And the prospect's data stays in your system from the first form submission. No third-party intake service holding your pipeline, no export step where attachments go missing. The lead's whole journey — inquiry, synopsis, conflict check, consultation, retainer, matter — happens inside the same governed platform that will run the case.