The Problem: Law Firms Lose Leads in the Intake Gap
Most small law firms have the same invisible problem. A potential client fills out a contact form at 11 PM on a Tuesday. By Wednesday morning, that lead has already called two other firms. If your intake process runs on manual effort — someone has to read the form, evaluate the case, draft a response, send a retainer, create a matter in Clio — you are consistently losing clients to firms that move faster.
This is the intake gap. It exists not because firms are careless, but because intake is genuinely complex work. You need to qualify the lead, assess the case type and potential value, send the right engagement documents, collect retainer payment, and open the matter — all before a single billable minute is logged. For a solo attorney or a small firm with one paralegal, that process can take 3–4 hours per qualified prospect.
The system I built for this client eliminates that gap entirely. From form submission to an attorney-notified, fully-funded Clio matter: under 5 minutes, zero manual steps for standard cases.
What Was Built: A 4-Stage AI Intake Pipeline
The entire system runs in n8n, triggered by a form submission. It uses Claude as the AI reasoning layer and integrates with Clio (the legal practice management platform most common in small US firms), DocuSign for engagement letter delivery and e-signature, and Stripe for retainer collection.
Stage 1 — Lead Scoring and Case Qualification
When a potential client submits the intake form, Stage 1 activates within seconds. The form collects: case type, incident date, jurisdiction, a brief description of the facts, and contact details.
An n8n workflow passes this data to Claude with a structured prompt that evaluates four dimensions:
- Case viability: Is this a case type the firm handles? Are there obvious disqualifying factors — statute of limitations issues, jurisdiction mismatches, or case types outside the firm's practice areas?
- Potential case value: Based on the case type and facts described, what is the likely settlement or award range for this jurisdiction?
- Lead quality score: A 1–10 numerical score with a one-paragraph rationale
- Recommended action: Auto-advance to Stage 2, flag for attorney review, or send a polite rejection
Claude returns structured JSON with these four fields. The n8n Switch node routes the case accordingly. High-scoring cases (7 and above) advance automatically to Stage 2. Cases flagged for review are logged to a Google Sheet and a Slack message goes to the attorney with the AI's full assessment — the attorney can decide in under 30 seconds whether to advance or reject. Rejected cases receive a pre-written, polite response. All of this happens within 60 seconds of form submission — including at 11 PM on a Tuesday.
Stage 2 — Engagement Letter Generation
For cases that advance, Stage 2 drafts a customized engagement letter using the intake form data. Claude fills in the case-specific details — the type of representation, the specific facts as described by the prospect, the fee structure relevant to that case type — into a master template that the firm's attorney reviewed and approved once before the system went live.
The letter is generated as a PDF and sent to the prospect via DocuSign for e-signature. The prospect typically receives it within 90 seconds of submitting their intake form. That response speed is genuinely surprising to prospects and immediately differentiates the firm from competitors whose follow-up happens the next business morning.
The n8n workflow then waits for the DocuSign webhook to confirm signature before proceeding. If no signature is received within 48 hours, an automated follow-up sequence fires: a reminder email on hour 24, a gentle SMS nudge on hour 48.
Stage 3 — Retainer Collection
Once the engagement letter is signed, Stage 3 triggers immediately. A Stripe payment link for the retainer amount — configured per case type in the firm's workflow settings — is sent to the client via both email and SMS.
This stage removes the most awkward part of traditional intake: asking for money. The automation handles it at precisely the right moment — right after the client has committed by signing the engagement letter. Retainer conversion rates at this stage are measurably higher than when retainer requests are sent separately the next day, because there is no gap for buyer's remorse to develop.
n8n waits for the Stripe payment webhook before Stage 4 activates. Failed payments trigger a follow-up with alternative payment instructions.
Stage 4 — Clio Matter Creation and Attorney Notification
With the retainer paid, Stage 4 creates the matter in Clio via the Clio API. All client contact details, case information, the Claude-generated case summary from Stage 1, and the signed engagement letter PDF are attached to the matter automatically. The matter is assigned to the appropriate attorney based on case type routing rules configured in the workflow.
The attorney receives a Slack notification with: the matter number, client name, case type, the AI-generated case summary, and a direct link to the opened Clio matter. Their first interaction with a new case is to review an already-opened, fully-funded matter with all documentation in place — not to chase down a form submission, draft a letter, or follow up on a payment.
The Technical Architecture
The system runs entirely in n8n, self-hosted on a VPS for full data control. The node sequence:
- Webhook trigger — receives the intake form POST
- Claude AI node — lead scoring with structured JSON output
- Switch node — routes to auto-advance, attorney review queue, or rejection path
- Claude AI node — engagement letter content generation using firm template
- HTTP Request node (DocuSign API) — creates and sends the envelope
- Wait node — pauses until DocuSign webhook fires with signature confirmation
- HTTP Request node (Stripe API) — creates payment link, sends via email and SMS
- Wait node — pauses until Stripe webhook fires with payment confirmation
- HTTP Request node (Clio API) — creates matter with all fields and attachments
- Slack node — attorney notification with matter link and AI case summary
Each stage has error-handling branches. If the Clio API call fails, the attorney gets an immediate Slack alert so no case falls through the cracks even if a third-party API has an outage.
Results After 30 Days
- Average time from form submission to attorney notification: 4 minutes and 12 seconds for auto-advance cases
- Manual paralegal work per standard intake: Zero. Paralegal capacity fully freed from intake administration.
- Intake volume handled: 3× the previous monthly volume with the same staff — because paralegal time was no longer consumed by intake admin
- AI classification accuracy: 100% correct on jurisdiction mismatches and case type rejections in the first month. Zero false positives that sent rejection letters to viable cases.
What Made This Work — and What Required Iteration
The AI scoring prompt took significant tuning. The first version was too permissive — advancing cases with obvious disqualifying factors. We ran 35 historical intake forms through the system before going live and adjusted the prompt until the AI's classification matched the attorney's known decisions on all 35 cases. The final prompt includes explicit disqualifiers specific to this firm's jurisdiction and practice areas.
The engagement letter template was written by the attorney, not generated by AI from scratch. Claude fills in the case-specific details. The attorney spent two hours writing the master template once, and it now generates correctly — and legally defensibly — for every applicable case type. This is the right division of labor.
The fallback sequences matter as much as the main flow. If DocuSign does not receive a signature within 48 hours, a follow-up sequence fires. If the Stripe payment link is unused within 24 hours, an SMS reminder goes out. If the Clio API call fails, an immediate Slack alert goes to the attorney. None of these were in the first version — all were added after observing real usage in the first two weeks.
Adapting This for Other Firm Types
The 4-stage structure — qualify, engage, collect, open — is general. The integrations depend on what the firm uses. Clio is the most common platform for small US firms, but the same pattern works with MyCase, Smokeball, or PracticePanther via their APIs. DocuSign can be replaced with PandaDoc or Adobe Sign. Stripe can be replaced with LawPay. The scoring prompt needs to be customized per firm — what makes a good personal injury lead is different from what makes a good estate planning lead.
For a firm with existing tools and an attorney available to review the engagement letter template and scoring criteria, a full rebuild typically takes 8–10 hours of configuration plus 2–3 hours of attorney time for template review and prompt calibration.
Want This Built for Your Firm?
If your law firm's intake process still relies on someone manually reading forms, drafting responses, chasing payment, and opening matters — this system eliminates all of that. The upfront build is recovered quickly through reduced paralegal hours and improved retainer conversion rates from faster response times.
Book a free 20-minute consultation to discuss whether AI intake automation makes sense for your firm's practice areas, current tools, and intake volume. If it is not the right fit, I will tell you that too.