AI Build Plan
Prepared for Phil Gore · Bluebonnet Properties

Start with the deceased-owner queue. Then expand.

After today's working session, the plan is narrower and more useful: use your base tax, auction, code, probate, and county data to identify confirmed and high-likelihood deceased owners first. The portal becomes a deceased-owner verification workbench: evidence-backed queues, sortable confidence, stale-data refreshes, and a WhatsApp-accessible research agent that helps you work the list faster without forcing a CRM replacement.

Get Started → View Build Plan
Scroll
↓
Business Profile

Bluebonnet Properties

Texas-based tax delinquent investor focused on deceased-owner acquisitions. Current workflow is spreadsheet-first, with a lightweight internal lead workspace for complicated heirship and probate situations.

🏠
Strategy
Tax Delinquent · Deceased Owner Acquisition
📍
Markets
Texas Counties · Oklahoma
👥
Team
Phil + Haroon/Harin + spreadsheet workflow
📄
Starting Data Set
30,000+ records loaded; more counties in progress
🎯
Target Record Profile
Deceased Owner · 3 Heirs or Fewer · $100K+ Equity
🔍
Current Research Stack
Sheets · Lead workspace · County records · WhatsApp
Bottlenecks Identified

What today's feedback clarified

Phil's priority is a practical first version: work from the base data, find deceased owners accurately, keep records fresh, and organize complicated leads without forcing a full CRM.

1
Priority #1: confirmed and high-likelihood deceased-owner detection.
The first useful version should identify confirmed and high-likelihood deceased owners from the base dataset with a clear evidence trail. Find a Grave, Ancestry-led obituary trails, Legacy-style obituary sources, probate records, and affidavits of heirship become first-class evidence sources.
🔴 Critical
2
Base data first. Add PropStream later only after the verification engine proves itself.
Start with tax delinquent lists, pre-foreclosure/tax auction lists, code violations, and probate. Once the deceased-owner queue is producing good results, Phil has PropStream lists that can be run through the same engine.
🔴 Critical
3
Records get stale. Tax status and amounts need refresh cadences.
Taxes get paid, amounts change, and stale data wastes research time. The plan now includes source/county refresh metadata, monthly re-scrubs where practical, and quarterly-or-better base-list refreshes depending on each county's difficulty.
🔴 Critical
4
County interfaces should become reusable adapters, not one-off scrapers forever.
Dallas is unique, Tarrant-style tax pages are common, and Houston has its own pattern. We will classify each county by interface type so the refresh/scrape process gets easier over time instead of starting from scratch each time.
🟡 High
5
WhatsApp is the first agent endpoint.
Phil said WhatsApp is probably the main work channel. The first connection page should let him connect WhatsApp and talk to the agent there: check an owner, explain a deceased score, summarize evidence, or run deeper research on a lead.
🟢 Medium
6
Support the dashboard/workspace Phil already wants, without forcing a CRM.
Phil's team lives in spreadsheets and a lightweight workspace for complicated leads. The portal should preserve that: statuses/tags, lead detail pages, heirship tree context, Zillow/probate/tax-suit/affidavit attachments, phone numbers, and notes in one place.
🟢 Medium
Build Plan

What we're building

Updated after today's call: v1 is a focused deceased-owner verification workbench, not a generic full CRM. The full stack supports the narrow outcome first, then expands once the queue is working.

Deceased-Owner Workbench Dashboard
Phase 1
A purpose-built workbench for the base dataset. County/source filters, sortable deceased confidence, evidence counts, last-refreshed dates, deal priority, and a record detail card that explains why a lead is confirmed, high-likelihood, possible, no-evidence, or needs review.
Timeline Immediate v1
Impact Gives Phil the money queue first: confirmed and high-likelihood deceased-owner leads.
Base Data + Refresh Cadence Layer
Phase 1
Track each county/source with last pull, last re-scrub, refresh method, interface type, and notes. Tax delinquent amounts/statuses should be re-scrubbed roughly every 30 days where practical, with base-list refreshes set county by county.
Timeline Immediate v1
Impact Keeps the list from going stale and makes county refresh work repeatable.
Evidence-Backed Death Verification Engine
Phase 2
The core v1 engine. Multi-source death verification across Find a Grave, Ancestry/Legacy-style obituary trails, probate filings, affidavits of heirship, county records, and other obituary sources. Each result must show source links, confidence, and a human-review reason when uncertain.
Timeline Immediate v1
Impact Turns the base dataset into a workable confirmed/high-likelihood deceased-owner queue.
Separate Death Confidence from Deal Priority
Phase 2
Two scores instead of one opaque number: death confidence explains the evidence that the owner is deceased; deal priority uses tax amount, equity, probate/code/pre-foreclosure signals, and lead quality. Sorting must work cleanly by both.
Timeline Weeks 4–5
Impact Prevents death evidence and deal economics from getting mashed into one confusing score.
Lead Workspace + Attachment Context
Phase 2
Mirror the workflow Phil showed: when a lead gets complicated, it needs one workspace for the heirship tree, Zillow photos, probate files, tax suit files, affidavits of heirship, phone numbers, status tags, and notes.
Timeline Weeks 4–5
Impact Keeps complicated opportunities from getting lost between spreadsheets and separate files.
WhatsApp-Connected Research Agent
Phase 2
Phil can talk to the agent from WhatsApp: check this owner, explain this death score, summarize the evidence, research this heirship, or run a deeper pass on a lead. Dashboard chat can come later; WhatsApp is the first practical endpoint.
Timeline Weeks 5–6
Impact Meets Phil where he actually works instead of making the dashboard the only interface.
Spreadsheet-Friendly Status + Queue Tracking
Phase 2
Simple statuses/tags and queues that can work alongside spreadsheets: confirmed deceased, high-likelihood, possible, no evidence, needs review, active lead, researching heirship, contacted, deal stage. No forced CRM migration.
Timeline Weeks 5–6
Impact Keeps leads from being left behind while preserving Phil's spreadsheet-first workflow.
Later Expansion: PropStream + Additional Lists
Phase 3
After the deceased-owner engine is working on the starting data, Phil can feed PropStream lists and additional lead sources through the same verification process. Expansion comes after v1 proves accuracy.
Timeline Weeks 5–6 + Ongoing
Impact Lets the system expand without distracting from the first outcome.
Timeline

Focused v1 first. Full expansion after the queue works.

Today's direction is clear: ship the deceased-owner verification queue first, connect WhatsApp, then expand sources and workflow depth once Phil can work the list.

V1 — Useful First Version
Base data + deceased-owner queue + evidence details
1
Base dataset loaded and normalized
  • Current Bluebonnet/Lead Forge records visible
  • Tax delinquent, pre-foreclosure/tax auction, code violation, and probate sources labeled
  • County/source refresh metadata added
2
Sortable verification queue live
  • Death confidence, deal priority, evidence count, status, and last refresh sortable
  • Record detail card explains each verification result
  • Human-review bucket catches ambiguous matches
3
WhatsApp + workspace pass
  • WhatsApp connection page for Phil
  • Lead workspace mirrors Phil's real workflow
  • Phil tests confirmed/high-likelihood queue
Next — Expansion After V1 Works
More sources, deeper heirship, PropStream, and automation
4
County adapter expansion
  • Dallas, Tarrant-style, Houston-style, and generic county tax interfaces classified
  • 30-day re-scrub cadence where practical
  • Quarterly-or-better base-list refreshes by county difficulty
5
PropStream and additional list ingestion
  • Run PropStream lists through the proven verification engine
  • Add deeper genealogy and heirship automation after v1 is accurate
  • Keep WhatsApp as the first live agent interface
Later — CRM/workflow depth only if needed
Add more CRM-like functionality only where Phil's spreadsheet/workspace workflow actually needs it. The goal is not to force a new CRM; it is to stop high-value deceased-owner leads from getting lost.
Ongoing — Monthly build credits
Additional counties, PropStream lists, more evidence sources, and workflow automations are added after the confirmed/high-likelihood queue is useful.
Out of Scope

What this is not

Clear boundaries. No surprises.

Investment

Simple pricing. One build. One monthly.

You're getting a system built on infrastructure that took two years to build for Dallas. You're not starting from scratch.

One-Time Build
$5,000
Focused v1 deceased-owner workbench, evidence-backed verification, refresh cadences, WhatsApp agent endpoint, and expansion path.
At Signing
$5,000
Payment link provided
Build Duration
~6 wks
No second payment
Pay $5,000 to Get Started →
Payment link sent after you confirm. Build kicks off same week.
After the build — $1,000/month ongoing
$1,000/mo
Ongoing Support
Developer hours each month. Add counties, ingest PropStream lists, maintain county adapters, improve death-verification accuracy, and add workflow features only after they prove useful.
Next Steps

How we start

Next step is alignment on the focused v1: deceased-owner queue first, WhatsApp endpoint, and refresh cadences baked in.

RE Reset
This document is private.
Enter the password to continue.