From bb93ba36881853ea07aba6500fbadac4379116ee Mon Sep 17 00:00:00 2001 From: Tony Balascio Date: Wed, 12 Aug 2026 23:12:22 +0000 Subject: [PATCH] docs: update README for VeriPath audits and clean data focus --- README.md | 183 +++++++++++++++++------------------------------------- 1 file changed, 58 insertions(+), 125 deletions(-) diff --git a/README.md b/README.md index 8d9851b..c0c6172 100644 --- a/README.md +++ b/README.md @@ -1,37 +1,27 @@ -# Digital Operations Partner +# VeriPath -A human-supervised, agent-assisted digital operations service for local service businesses. +A human-supervised verification service for local businesses. We prevent silent customer loss by ensuring customers can find, trust, and engage a business, and that search engines and AI systems can find, understand, and accurately represent it. -This project exists to prevent silent customer loss by making sure customers can find, trust, and engage a business, and that search engines and AI systems can find, understand, and accurately represent it — often before the business would notice something was wrong. - -This repository defines the business foundation, operating model, evidence standards, client lifecycle, Path to PoC sequencing, and tooling for a Digital Operations Partner service. - -**New hire start here:** `docs/operations/employee-onboarding-manual-v1.md` +This repository defines the audit methodology, evidence standards, canonical business record schema, client records, and operational workflow for VeriPath. --- -## Now vs Later (read this first) +## Current Focus (read this first) | Horizon | What it is | Status | |---------|------------|--------| -| **NOW — Operating mode** | Human-run cold audits (GBP form → Leonard), Layer 1a public inventory, Layer 2 threats, Path to PoC → economics → retainer | **Active process** | -| **LATER — Product horizon** | Public Assessment AI, fixed-fee gates, State Clock, Sentry, Market Radar, multi-tier product packaging | **Not operating process** — strategy only | +| **NOW — Operating mode** | Manual audits: evidence capture, AI representation testing, Silent Customer Loss Maps, canonical business records, owner review | **Active** | +| **LATER — Deferred** | GeoLocal integration, MCP/agent infrastructure, payment flows, automated multi-tenant platforms | **Deferred** | -- **Operating truth for day-to-day work:** this README, Path to PoC, Audit Playbook, Agent Charter, onboarding manual, templates, tools. -- **Strategy / thesis (do not treat as current SOP):** `docs/strategic-frame.md`, `docs/whitepaper.md` +**Operating truth:** this README, evidence standards, canonical record schema, audit methodology, and client records. -If a later-horizon idea conflicts with the operating docs, the operating docs win until explicitly promoted with a dated decision. +If a deferred item conflicts with current operating docs, the operating docs win until explicitly promoted with a dated decision. --- ## Working Definition -> We continuously prevent silent customer loss for local businesses by making sure customers can find, trust, and engage the business, and that search engines and AI systems can find, understand, and accurately represent it, often before the business would ever notice something was wrong. - -This is a working definition, not a final brand statement. - -For the full business requirements, see: -`docs/business-model/business-requirements.md` +> We continuously prevent silent customer loss for local businesses by making sure customers can find, trust, and engage the business, and that search engines and AI systems can find, understand, and accurately represent it, often before the business would notice something was wrong. --- @@ -42,7 +32,6 @@ For the full business requirements, see: Silent Customer Loss is the umbrella term for customer acquisition problems that occur without the business owner clearly seeing the loss. It includes: - - **Customer Path Leakage** — a customer has interest or intent, but digital friction prevents contact, booking, visiting, or engagement. - **Discovery Failure** — a potential customer never meaningfully encounters the business because customers, search engines, or AI systems do not find, understand, surface, or accurately represent it. @@ -54,12 +43,15 @@ It includes: ### Evidence Model -- **Tier 1: Verified Evidence** — concrete, timestamped, observable, or instrument-backed findings. -- **Tier 2: Indicative Evidence** — sampled, directional, probabilistic, or risk-based findings. +- **Tier 1: Verified** — concrete, timestamped, observable, or instrument-backed findings, confirmed by owner or independent sources. +- **Tier 2: Corroborated** — confirmed by independent sources. +- **Tier 3: Indicative** — single source, probabilistic, or directional. -Agents may propose classifications; humans approve final tiers before client-facing use in Version 1. - -Competitor facts about third parties are **Tier 2 by default** and never the primary claim of a threat (see Audit Playbook). +Field states: +- `observed` — captured from a source but not independently confirmed +- `verified` — confirmed by independent sources or owner approval +- `stale` — previously accurate but freshness has expired +- `disputed` — conflicting evidence exists; resolution pending ### Operating Rule @@ -67,43 +59,40 @@ Competitor facts about third parties are **Tier 2 by default** and never the pri --- -## Path to PoC (Governing Sequence) +## Audit Workflow -Master sequence: `docs/architecture/path-to-poc-sequencing.md` - -Layers (fixed order): - -1. Account Intelligence — **1a Public baseline** (cold audit) / **1b Owner-access** (engagement) -2. Threat Diagnosis (Threat Register) -3. Economics & Pricing Validation -4. Requirements -5. Workflows -6. Agent Build & Test -7. POC Delivery - -Client artifacts live under `docs/clients//`. - -Cold audits can **complete Layer 1a** without owner credentials. Full engagement-grade Layer 1 requires 1b or explicit acceptance of Tier 2 ceiling. +``` +Intake +→ evidence capture (website, GBP, NAP, schema, listings) +→ AI representation test (buyer-intent queries across models) +→ Silent Customer Loss Map +→ evidence classification (tier + source independence) +→ draft canonical business record +→ owner review and confirmation +→ correction plan +→ before/after measurement +``` --- -## Agent & Audit Tooling +## Canonical Business Record -| Doc / Tool | Path | -|------------|------| -| **Employee Onboarding Manual** | `docs/operations/employee-onboarding-manual-v1.md` | -| **Task reliability ledger** | `docs/operations/task-reliability-ledger.md` | -| Agent Charter | `docs/agents/agent-charter-v1.md` | -| Audit Playbook | `docs/agents/audit-playbook-v1.md` | -| Leonard GBP Execution Card | `docs/agents/leonard-gbp-execution-card.md` | -| GBP Snapshot Intake (markdown) | `docs/templates/gbp-snapshot-intake.md` | -| GBP Snapshot Form (HTML) | `tools/gbp-snapshot-form.html` | -| Data Inventory template | `docs/templates/data-inventory-template.md` | -| Threat Register template | `docs/templates/threat-register-template.md` | -| Requirements template | `docs/templates/requirements-template.md` | -| Workflows template | `docs/templates/workflows-template.md` | +Schema: `docs/architecture/canonical-business-record-v0.1.1.md` +Policy: `docs/architecture/evidence-freshness-policy-v0.1.md` -**Cold audit GBP workflow:** open `tools/gbp-snapshot-form.html` → fill → generate → paste to Leonard with the execution card instruction. Score runs in the task reliability ledger. +VeriPath is the system of record. Downstream consumers (GeoLocal, APIs, MCP tools) treat the canonical record as read-only. + +--- + +## Decision Records + +Core philosophy and architectural decisions are recorded in `docs/decisions/`. See the roadmap for the current execution plan. + +| Decision | Path | +|----------|------| +| Scope narrowed: VeriPath first, GeoLocal deferred | `docs/decisions/2026-08-13-narrowing-scope-veripath-first.md` | +| VeriPath owns canonical truth | `docs/decisions/2026-08-12-veripath-system-of-record-boundary.md` | +| Execution roadmap v0.1 | `docs/roadmap/execution-roadmap-v0.1.md` | --- @@ -111,30 +100,8 @@ Cold audits can **complete Layer 1a** without owner credentials. Full engagement | Client | Status | Path | |--------|--------|------| -| Phoenix Salon + Spa | **Active** — Layer 1a Cold Audit (Pilot) | `docs/clients/phoenix-salons/` — data: `client-data-pss` | -| Overcome Fitness | Draft Benchmark (Layer 1a path; GBP still open; not engagement-complete) | `docs/clients/overcome-fitness/` | - ---- - - - -## V1 Scope Exclusions - -Out of scope for V1 unless separately quoted: - -- Social media production -- Paid advertising -- Website redesign -- Branding projects -- Content marketing -- Full SEO campaigns -- CRM / email marketing - ---- - -## Go-To-Market Motion - -Proof-of-concept first. Demonstrate value through focused assessment, evidence-backed findings, a low-risk PoC fix when appropriate, and a before/after report. Retainer follows demonstrated value. +| Phoenix Salon + Spa | Active — audit and canonical record | `docs/clients/phoenix-salons/` — data: `client-data-pss` | +| Overcome Fitness | Draft benchmark | `docs/clients/overcome-fitness/` | --- @@ -142,55 +109,21 @@ Proof-of-concept first. Demonstrate value through focused assessment, evidence-b ``` docs/ - vision/ - architecture/ # Path to PoC, economics guide, principles - business-model/ - use-cases/ - templates/ # Layer templates + GBP intake - agents/ # Charter, playbook, execution cards - clients/ # Per-client instances - operations/ # Employee manual, reliability ledger - workflows/ - roadmap/ - strategic-frame.md # LATER horizon (not SOP) - whitepaper.md # LATER horizon (not SOP) -research/ -specifications/ -implementation/ -tools/ # Operational forms (GBP snapshot HTML, etc.) + decisions/ # dated decision records + roadmap/ # execution plan and phases + architecture/ # schema, evidence policy, agent boundaries + clients/ # per-client audit records and reports + operations/ # operational manuals and ledgers +research/ # competitive and market research +specifications/ # JSON schemas and data contracts +implementation/ # automation scripts and tools +tools/ # operational forms (GBP snapshot, etc.) ``` --- -## Current Phase - -**Phase: Path-to-PoC execution + first client benchmark (operating mode)** - -Resolved foundations: - -- Mission and working definition -- Evidence model and operating rule -- Client lifecycle and V1 exclusions -- Path to PoC sequencing (Layer 1a / 1b split) -- Economics & pricing execution guide -- Audit playbook + Layer templates + competitor evidence rules -- GBP cold-audit intake form + Leonard execution card -- Employee onboarding manual -- Task reliability ledger (empty counters until first scored runs) -- Charter amended: validation client is any real Layer 1–3 data path (Overcome exception logged) - -Open work: - -- GBP snapshot for Overcome Fitness (human form pass) before 1a lock -- Economics validation with real client numbers -- Leonard scored runs recorded in the ledger -- Pricing validation and delivery packaging -- Optional: path manifest / doc validator - ---- - ## Philosophy -Users → Problems → Requirements → Workflows → Agents → Tools +Users → Problems → Requirements → Evidence → Records → Automation -No automation runs until the business is understood. +No automation runs until the manual workflow is proven. No infrastructure is built until the audit produces paid value.