Legal Intake Studio
Estate-planning intake assistant: paralegal walks the client through a branching question flow, conflict flags surface in real time as inconsistencies are entered (special-needs SSI disqualification, beneficiary-direct-conflict, fiduciary-mismatch), and the studio generates two side-by-side documents — a paralegal action draft and a senior-attorney review summary. Send to the senior-attorney review queue for approve / revise / annotate. Audit log captures every action. Built for estate-planning, transfers cleanly to family / PI / immigration / business-formation intake.
How it works
1. Pick a scenario, walk the interview
Three pre-loaded estate-planning scenarios — married couple with two minors (Hollister), single physician practice owner (Krishnan), blended family with a special-needs son on SSI/Medicaid (Vega). Branching question flow with conditional logic (married → prenup, business owner → succession), inline ‘why we ask’ help, progress rail with section completion + cumulative flag count.
2. Conflict flags fire inline as answers come in
A pure-function conflict-detection engine evaluates declarative rules against the live answer state on every keystroke. Three severity tiers: gold (informational), amber (review needed), red (blocking — like a direct $200k bequest to a special-needs son that would disqualify him from SSI/Medicaid, or a minor child named as direct 50% brokerage beneficiary without a UTMA / 2503(c) trustee). Click any flag to see both conflicting answers, why it matters legally, and the suggested resolution.
3. Documents stream + senior review queue
Hit ‘Generate outputs’ — the paralegal action draft and senior-attorney review summary stream side-by-side paragraph-by-paragraph (~6-12s each). Paralegal draft has [ATTORNEY INPUT REQUIRED: …] markers and an open-items checklist; senior summary opens with ‘Items Requiring Your Decision’ ranked by severity and closes with concrete recommendations. Send for senior review → card lands in the queue → senior approves / requests revision / annotates → audit log captures the full chain.
60-second Loom
75-second flow: scenario picker → click Hollister (married + 2 minors) → walk the questions → answer ‘inherited shotgun’ → amber firearms-disclosure-mismatch flag fires inline → click flag, side panel shows both conflicting answers + reason + resolution → ‘children directly 50/50’ on brokerage beneficiary → red blocking flag fires (minor + no UTMA/trust) → DPOA: spouse explicitly excluded → amber fiduciary-mismatch flag fires (spouse-as-executor + spouse-excluded-from-DPOA) → progress rail shows 1 blocking + 2 review → Generate outputs → both documents stream paragraph-by-paragraph with [ATTORNEY INPUT REQUIRED: …] chips → Send for senior review → Approve → status flips to ‘Approved by Eleanor Vance’ → audit log shows the full chain.
Get the full build spec PDF
One-page PDF with the architecture, the tech stack, the timeline (2–3 weeks), and the investment range ($9,990 – $25,000). Sent to your inbox in under a minute.
Who it’s for
Estate-planning, family-law, PI, immigration, and small-business-formation firms — managing partners and ops leads who need junior attorneys and paralegals to capture intake cleanly, surface conflicts before they become drafting errors, and hand a senior partner a 1-page review summary instead of 40 pages of raw notes.
Tech under the hood
Frequently asked questions
Is this real legal advice?
No — every page carries a ‘demo only — not legal advice’ notice. The scenarios are fictional, the conflict rules are illustrative of common estate-planning intake failure modes, and document templates are written to be educational, not authoritative.
Where does the LLM live?
The demo streams pre-written paragraphs with a typewriter cadence so the UI feels live without a real LLM call (zero credentials, zero cost, deterministic playback). In a real engagement we wire the document generation to Claude or GPT with a server-side proxy, retrieval over the firm’s prior matters, and a redaction layer for PII before any external call.
How are conflict rules authored?
As declarative TypeScript objects in `data/seed.ts` — each rule has an `involves` list (the question IDs it depends on), a pure-function `match(answers)` predicate, a severity level (info / amber / red), and a resolution-guidance payload. A paralegal can author new rules without touching React. The engine re-runs on every answer change and only fires rules whose involved questions are fully answered (so no premature flags on partial data).
Does it integrate with Clio / MyCase / NetDocuments?
Not in the demo — out of scope per the brief. In production we wire matter-creation to Clio Manage or MyCase via REST, drop generated documents into NetDocuments or a Microsoft 365 / SharePoint matter folder via Graph API, and pull engagement-letter templates from the firm’s document library. Adding a new LMS adapter typically takes 3-5 days.
Want one of these in your business?
We’ll scope, build, and hand off a production-ready system in 3–4 weeks.