Zedric.Book
← All filesStatus: delivered · live

CS-13 // Dossier

Regulated-market lead engine

A complete lead-capture-to-review CRM build for a home-care agency that started from an empty account — ten cross-wired workflows, an AI intake agent, and a compliance architecture that makes it structurally impossible to message a referral coordinator.

Exhibit // CS-13

R-02 · interface redrawn from the live build — zero client pixels

Redrawn exhibit — intake console: four entry channels feeding a step-one tag guard that ends the run on a coordinator tag, a ten-workflow register totaling 131 steps, and a zero-messages-to-coordinators stat
Client
R-01SEALED UNDER NDA. SPECIFICS WALK THROUGH LIVE. → BOOK
Industry
Home-care agency · multi-county footprint, three payer lanes
System
Regulated-market lead engine
Stack
GoHighLevel · Conversation AI · A2P SMS · Playwright
Trigger
An empty account, and a referral channel where one wrong message is a felony
Metrics
System scope

01 // The problem

The account was effectively empty — zero calendars, a stock pipeline holding demo records, no capture path. And the agency’s proven referral channel is state-program referral coordinators, where anti-kickback law makes sending anything of value to a coordinator a crime — so the automation could not be allowed to message them at all.

02 // The system

Redrawn — no client data

The system

FOUR ENTRY DOORSCALL · SMS · CHAT · FORMCOORDINATOR GUARDENDS THE RUNSPEED-TO-LEAD+ OPPORTUNITY CREATEDAI INTAKE AGENTQUALIFIES + BOOKSCONSULT BOOKEDINTERNAL TASK ONLYZERO SENDS

03 // The delta

Before and after go-live

Before

  • Zero calendars, zero capture path, a stock demo pipeline
  • The one proven referral channel legally untouchable by automation
  • No tag taxonomy, no custom fields, no lost reasons
  • A third-party audit with false findings about to be actioned

With the system

  • Ten live workflows, 131 steps, four entry doors covered
  • Messaging a coordinator is structurally impossible, audited
  • A 23-tag namespaced taxonomy and full field architecture
  • The audit re-verified: four findings killed, four real gaps surfaced

04 // What got done

6 entries

  1. B-01Captured all eleven account surfaces read-only before designing anything, so the build started from live state, not assumption
  2. B-02Made the felony-risk constraint structural, not policy: a coordinator-tag guard that ends the run sits at step one of nine workflows; the tenth carries the inverse guard and zero outbound sends
  3. B-03Reverse-engineered the platform’s form renderer to recover the real field-mapping schema, and proved the fix with a browser accessibility snapshot rather than a 200 response
  4. B-04Documented fourteen platform substitutions where the API cannot express a design condition — each with a sanctioned fallback and the exact UI step to restore full behavior
  5. B-05Refused to guess on the setting that mattered most: read the engine’s own dispatcher to confirm no quiet-hours API existed, and escalated it as a human task instead of shipping a value that would look applied while doing nothing
  6. B-06Re-verified a third-party audit line by line against live reads — four findings proven false, two diagnoses backwards, and four launch-blocking issues it missed surfaced

05 // Outcome

System scope — capability metrics, honestly labeled

Scope
workflows · 131 steps, published live
0
Scope
messages possible to a referral coordinator — by structure
0
Scope
account surfaces captured read-only before any write
0

06 // The result

From the build log

  • All ten workflows published live, with the coordinator guard audited across every one
  • A calendar built from zero, live free-slot check passing
  • A research-grounded 90-day marketing plan built from the owner’s real numbers, not invented ones

Want a system like this one?

Free · 30 min · leave with a rough scope either way

Book the intro call