# geolocal.io — Business Plan (working draft)

**Branch:** `planning-selfserve-audit` (review packaging on `review/ready`)  
**Status:** In progress — conversation checkpoint, not final  
**Purpose:** **Only** durable home for the business idea on this branch. Fold strategy here; do not spawn sibling strategy docs.

**How we work this file**
- Checkpoint after solid decisions (or when a phrase must not be lost).
- Prefer revising this over spawning parallel strategy docs.
- Mark unknowns as open questions. Do not invent market numbers here.
- Nothing is locked until Ty says so.
- Scribe the **essence** of agreed direction. Do not promote open packaging into POR until it sticks.
- **Scribe:** SEO & AEO Desk (sole).

---

## 1. North star (wording still open)

**Intent (stable):** Enable local businesses to be discovered and recommended by AI assistants.

Still wordsmithing the exact sentence. The job does not change while the phrasing does.

---

## 2. What the company is

Infrastructure — not a consumer destination.

- Host structured business truth so AI assistants can discover, understand, and recommend local businesses (services and local retail with real-world presence).
- Give the business a simple way to point at that truth (no servers for them to run).
- Orchestrate booking / payment / related partners where needed; do not rebuild Calendly or Shopify.
- Stay invisible to end consumers; win by being useful to agents and trustworthy to owners.

**Not:** another Yelp/Google for human browsing attention.

---

## 3. Who we serve

**Primary customer:** the business owner.

By extension:
- Someone in their organization acting for them
- A partner they already trust (agency, SEO, chamber, tourism board)

Partners are a **path to the owner**, not a different product.

---

## 4. How demand shows up (two moments)

1. **Presence for other AIs** — When a consumer’s assistant asks for a local plumber / salon / store, structured geolocal data can be consulted and acted on.
2. **Owner (or their AI) asks for help** — “How do we help our business be discovered by AI?” Invest in scenarios where the owner’s own AI can diagnose and set that up — not only human sales through agencies.

Same north star; different moments of truth.

---

## 5. Wedge and flywheel (direction agreed)

**Preferred wedge:** a free AI-visibility checkup.

Why:
- Shows immediate value / need
- Builds critical mass
- Small free actions amplify adoption

**Long-term reputation bet:** become so useful and dense that AI systems default to checking geolocal for local-class queries — analogous to how they lean on Yelp or Foursquare today.

**Strategic read:** free usefulness → coverage → agent trust → more local queries resolved via geolocal → more owners want in → denser graph.

Evidence (category signal, not brand list): [`docs/evidence/SBE-TECH-USE-SURVEY-2026.md`](./evidence/SBE-TECH-USE-SURVEY-2026.md) — high AI tool use among SMBs; research/content dominate; local discovery gap; peer/WOM trust path.

---

## 6. Owner experience

### 6.1 Initial workflow (working direction)

1. Arrive on an **outcome promise** (not plumbing)
2. Tell us the business
3. **Validation gate** — prove you represent the business
4. Dual-scoreboard audit in owner language + ranked foundation gaps
5. Short punch list (3–7 actions)
6. Optional durable step (minimal owned geolocal record on free path)
7. Return hook for re-check

### 6.2 Surfaces (working direction)

- **V1:** `geolocal.io/getstarted` (or similar) with **inline AI chat** driving intake — not forms-first.
- **Parallel from day one (design):** owner-agent → Owner MCP (same workflow).
- Thin status page behind login can hold the scorecard later.
- Web chat = on-ramp + control; agent-to-agent = scale + distribution.

**Hard rule:** chat and MCP are **clients of one backend workflow** (verify → dual-score audit → punch list). If they diverge, we’ve built two products.

### 6.3 Entry / discoverability angles (working — capture, not all build now)

Partner paths, search, owner’s own AI asking “how do we get discovered,” later marketplace listings — prioritize by trust (peers/WOM) and where the discovery question is asked.

### 6.4 Phase 1 conversation tone

Speak **outcomes**, not plumbing. Example: “most recommended bakery on Whidbey Island for muffins.” Do not lead with MCP, endpoints, or `llms.txt`.

---

## 7. Scorecard and tiers

### 7.1 Owner-facing scorecard (working — not locked)

Three top-line outcomes (wording TBD), each with plain “why not yet”:

1. **AI knows you exist** — findable identity / NAP / listings  
2. **AI knows what you do** — services, specialties, hours  
3. **AI would recommend you** — trust + fit for a real job in a place  

Not a 40-row SEO audit. Dual scoreboard: hygiene feeds rungs 1–2; recommendability is differentiation. How we *measure* “would recommend” (heuristics vs live assistant checks) stays open.

### 7.2 Discovery foundation + AI readiness

**Two tiers, one product story** (owner language):

1. **Discovery foundation** — Core listings + NAP/hours/category/contact consistency (internal: “2017”)
2. **AI readiness** — known-for queries, scorecard, structured presence when ready (internal: “2027”)

Measure both early; *fix* foundation first when broken. Skip invented hygiene if already clean.

**Owner vocab:** Core listings; The Core 3 (GBP, Apple Business Connect, Bing Places).

**Foundation bar** (brick-and-mortar local) — not reviews, photo perfection, or ranking #1:

1. One clear identity — correct NAP; no competing duplicates  
2. Claimed Core 3 (genre/market adds are separate)  
3. Accurate hours  
4. Right category + short plain “what you do”  
5. Reachable — working phone and/or website/contact  
6. Not contradictory — site vs Core 3 agree on name/place/hours  

**Foundation sync (P1)** across Core 3: NAP + primary category + hours + website URL + short description + service-area/address mode. P2 (attributes, photos, menus, booking, secondary categories) is genre/known-for gated.

### 7.3 Surfacing work — Layer A / B / C

Not one giant schema. **A** universal Discovery foundation → **B** thin genre pack → **C** known-for overlay. Scorecard first, then 3–7 actions max by impact.

Index + packs: [`docs/product/genre-packs.md`](./product/genre-packs.md).

### 7.4 Beyond NAP — provider capability ladder

*(Absorbed from former `LAYERS-BEYOND-NAP.md` — do not re-spawn that file.)*

Distinct from Layer A/B/C (what work we show). This is what providers can hold:

1. **Shared listing basics (Core 3)** — category, hours, website, description, service-area mode  
2. **Provider listing extras** — GBP services/posts/actions/attrs; Apple Good to Know / hours types; Bing thinner  
3. **Provider stacks** — Apple Stack (brand → Maps+Wallet/Pay, Showcases, Actions, Insights); Google = rich but fragmented listing + adjacent products  

**Why owners care:** NAP = findable; extras = know what you do; stacks/actions = chosen and acted on.

Field SoT: [`docs/product/genre-packs/core3-field-map.md`](./product/genre-packs/core3-field-map.md). Sources: [`SOURCES.md`](./product/genre-packs/SOURCES.md).

**Later proof layer (not free first-hour):** platform KPI telemetry (e.g. Apple Insights). Show only what platforms return — never invent numbers. Fit = ongoing AI readiness / retention.

### 7.5 Validation gate (essence)

Anyone using GeoLocal tools must pass a **business-representation check** first. Mechanism and free/silver/gold allotments stay open until COGS is real.

### 7.6 Free path after checkup (not locked)

1. Ranked DIY punch list  
2. Mint of a **minimal owned geolocal record** — densify does not wait for paid  

Soft upsell: “comfortable DIY, or want GeoLocal to help?”

**Bundles (free/silver/gold):** pinned open — capture value-prop angles first, then revisit.

### 7.7 Paid “help” vs product (not locked)

- **Paid listing cleanup** = partner / agency **fulfillment**, not the product.  
- **Product** = owned geolocal record + ongoing AI-readiness score (and later agent-facing presence).  

Do not document GeoLocal as a local SEO shop with extra steps.

---

## 8. Distribution investment (working direction)

**Invest in one portable Owner MCP** wired to the geolocal.io backend (same workflow as chat). Thin adapters/listings second — not five one-off apps.

| Priority | Surface | Call |
| :--- | :--- | :--- |
| P0 | Owner MCP + web chat | Build / debug privately first |
| P1 | GrokBot (dogfood) → Claude Connectors / ChatGPT plugins / Grok connectors | Thin listings on the **same** MCP |
| P2 | Square OAuth (optional) | Only if it shortens verify/NAP — not a second product |
| P3 | Square App Marketplace listing | **Defer** until free-checkup proof; marketplace onboarding without Partnerships contact was paused (2026) |

**Debug sequence:** private/local MCP → Ty + GrokBot → harden auth/rate limits → public listings. Remote connectors need a public MCP URL when opened.

**Anti-pattern:** separate “GeoLocal for Square / GPT / Claude” codebases.

---

## 9. Business plan audience (still open)

Primary reader for v1 structure: internal / investors / partners — pick one first.

---

## 10. Inputs (not on this branch)

`main` holds older planning (north star, canonical strategy, SSP use cases, etc.). Treat as **raw material only** — do not re-import into this branch. Resolve conflicts on purpose in this file when pulling an idea over.

Evidence and field sources that *are* on this branch live under `docs/evidence/` and `docs/product/genre-packs/SOURCES.md`.

---

## 11. Checkpoint log (trim, don’t spawn)

| When | What landed |
| :--- | :--- |
| 2026-09-15 | North star; owner-first; partner path; owner-AI “get discovered”; free checkup wedge; reputation flywheel |
| 2026-09-15 | Dual scoreboard; Discovery foundation → AI readiness; Core 3 vocab; foundation six-point bar; P1 sync fields |
| 2026-09-15 | Layer A/B/C; sourced Core 3 field map; salon/bakery application notes; Apple API names BLOCKED when gated |
| 2026-09-15 | Validation gate essence; free DIY + minimal record; paid cleanup = fulfillment; bundles pinned open |
| 2026-09-15 | Owner workflow + V1 chat + Owner MCP same backend; distribution: one MCP, thin AI adapters, defer Square Marketplace |
| 2026-09-15 | Unslop: locked placement map; absorbed beyond-NAP into this file; evidence/ folder; Desk = sole scribe |

---

## Next conversation topics

1. Wordsmith scorecard rungs + plain “why not yet” evidence  
2. How to measure “would recommend” honestly  
3. Primary reader for v1 of this plan  
4. Offer / packaging after value-prop angles captured  
5. First 3–5 Owner MCP tools (owner-facing)  
6. 90-day proof: what “AI checks us” requires in coverage/quality