diff --git a/docs/ALTERNATE_BUSINESS_PROPOSAL.md b/docs/ALTERNATE_BUSINESS_PROPOSAL.md index 16ed30b..b7dce39 100644 --- a/docs/ALTERNATE_BUSINESS_PROPOSAL.md +++ b/docs/ALTERNATE_BUSINESS_PROPOSAL.md @@ -117,13 +117,16 @@ Longer-term destination in the concept SoT (not version one): an industry graph ### Three claims (read this first) -**1. What the problem is.** +#### 1. What the problem is. + Picture a bakery owner who keeps losing easy customers. People show up after closing because the hours online are wrong. Others call a disconnected number and never call back. A few say they asked an assistant for a nearby bakery and got pointed somewhere else — or got her shop with the old phone number. She sits down on a Tuesday night with a clear job: fix this. She updates her Google listing — new phone number, new weekend hours — and thinks customers will find the right info everywhere. They will not. The same details still sit wrong or missing on Yelp, Bing, Apple Maps, and dozens of other online directories. A handful of big data companies quietly copy that stale information around the web. When someone asks an AI assistant — ChatGPT, Gemini, or anything like them — “where’s a good bakery near me?”, that assistant often pulls from those scattered, outdated copies. Sometimes the shop never shows up. Sometimes it shows up with last year’s number. The owner never sees the miss. Customers lose trust. Fixing one listing is not the same as being findable and accurately described wherever people discover local businesses today, including when they ask AI. -**2. What the industry looks like today.** +#### 2. What the industry looks like today. + Most of the “help your local business get found” industry still chases Google Maps wins: polish the Google business profile, collect reviews, post photos, clean up a few directory listings. That work can bring more phone calls from Google. Companies sell subscription tools that push your listing to many sites at once, score how “complete” your profile looks inside their own software, and track where you rank on a map. Almost none of those products answer the question owners now care about: what do AI assistants actually say about my business — do they find me, recommend me, and get the facts right? Outside research shows AI assistants name far fewer businesses than the short list of three Google shows at the top of many local searches, disagree with each other, and often get hours or phone numbers wrong even when they recognize the shop. People who work in this field warn that big listing subscriptions can lock up your data — stop paying and your listings may disappear. Small owners who do the work themselves usually get the credit for results. Customers are asking AI for local recommendations more than they did a year ago, but the tools for sale still mostly optimize last year’s game. -**3. What we can build to solve it — and why it can be a business.** +#### 3. What we can build to solve it — and why it can be a business. + We build a product that starts with what the owner can feel: ask the major AI assistants what they say about this business, then show what is missing, wrong, or stale. That check is the reason to sign up. Behind it, one action files corrections on the places that actually feed those answers — the big data companies that fan out into many directories, and the major sites people and machines trust (Google, Apple, Bing, Yelp, and similar) — so the owner does not fill out a dozen forms. We still help them keep Google strong, because Google is where many customers double-check after an AI suggestion, and because a correct Google listing is useful proof when we ask other sites to update. We measure success simply: after the fixes, ask the same AI questions again — does the business show up, and is the description accurate? That is a viable business because owners already pay for listing tools that do not answer that question, AI recommendations are rising fast, and the gap between “I fixed Google” and “AI gets me right” is painful, visible, and mostly unserved. ### Pain @@ -229,7 +232,7 @@ Classic local SEO is necessary infrastructure and the densest call evidence; AI ### Implications for this proposal -**Build toward:** +#### Build toward 1. **Accuracy-aware AI audit** — fixed buyer prompts; score Presence (found on claim surfaces / stomped copies) and Recommendability (named when it matters; facts match claimed truth). Prefer mentioned vs recommended vs **misstated** over vanity mention counts. 2. **One-click tickets for umbrellas (Lane A) and majors (Lane B)** — field / wrong / right / GBP-or-claimed-truth as proof. Umbrella-only leaves Core majors wrong; majors-only leaves Layer-1 feeds stomping the claim. @@ -239,4 +242,6 @@ Classic local SEO is necessary infrastructure and the densest call evidence; AI 6. **DIY result credit ⇒ product posture** — make tickets and ownership easy; avoid Yext-like hostage pipes where Presence evaporates when the subscription stops. Paid pipes remain relevant for scale (~30+ / 100+ locations per practitioner consensus), as optional scale lanes — not the identity of the product. 7. **Localeze / Foursquare honesty** — charge/time expectations (Localeze paywall, 30–90 day partners); Foursquare Placemaker suggest-edit is real; OpenAI adjacency is not a ranking claim; Managed Users later. -**Success check:** re-prompt the same asks; Presence and Recommendability improve (found, recommended, facts correct) — not “Localeze score went up,” not “we submitted 400 citations.” +#### Success check + +Re-prompt the same asks; Presence and Recommendability improve (found, recommended, facts correct) — not “Localeze score went up,” not “we submitted 400 citations.” diff --git a/docs/research-brief.pdf b/docs/research-brief.pdf index abacd2d..bd621d4 100644 Binary files a/docs/research-brief.pdf and b/docs/research-brief.pdf differ