Business Wiki — Overview

Properties1
tags

5 min read

Business Wiki — Overview

Evolving synthesis. Updated as sources are ingested and the thesis develops.

Sources ingested: 24 Last updated: 2026-09-04

Working Thesis

(Working hypothesis; 24 sources, most still secondary or vendor-authored.)

The most promising launch path for a Port Huron AI-automation consultancy is services-first, vertical, and local: sell manual AI-automation and Agent Readiness-style audit work to one or two Blue Water Area verticals now, accumulate proprietary local market data along the way (Niche Data Refinery), and productize the repeated work (Productized Services (services-first, productize later)). The Agent Web shift — with Cloudflare building payment rails like x402 / Pay-per-Crawl — suggests services and data built today could become metered agent-payable resources later, but the near-term revenue is in the services, not the rails.

Key Findings

  1. The internet is shifting from monetizing human attention to monetizing agent-usable resources; that favors clean data, structure, and vertical expertise (Agent Web).
  2. Sell to intermediaries first: agencies/consultants already serving a niche pay for market intelligence more easily than end business owners (Niche Data Refinery).
  3. “Sell the screenshot” — showing an owner what AI currently says (or doesn’t say) about their business — is a concrete, cheap-to-produce sales motion (Agent Readiness).
  4. The niche filter for data plays: valuable, repeatable, changing, fragmented, annoying.
  5. Services-first de-risks platform-timing bets while cash-flowing from day one.
  6. Verified: Cloudflare’s agent-payment stack is real but early — AI Crawl Control is GA, Pay per Crawl is private beta, the Monetization Gateway is waitlisted and stablecoin-only (Cloudflare). x402 is usable today inside Cloudflare’s Agents SDK and MCP servers, so paid agent tools are technically possible now even though general rails aren’t. This strengthens the services-first thesis: near-term revenue is in audits/automation, with agent-payable products as the later phase.
  7. Cloudflare’s broader Agents Week stack is more useful near-term as a delivery substrate than as a new product thesis: browser automation can produce mystery-shopper evidence; tracing/replay/approvals can make the proof-business workflow auditable; and WebMCP/AEO/AI Search can turn an AI-visibility finding into an agent-callable website remediation (Everything we launched during Agents Week, Agentic Service Operations).
  8. A standalone prompt-running AI-Visibility Audit is vulnerable to commoditization: Cloudflare now supplies readiness evidence and early-access AEO recommendation metrics. The local moat must be vertical interpretation, implementation, safe callable actions, and recurring before/after proof (From ranking to recommended: get your site ready to thrive in the age of AI agents, Agent Readiness).
  9. Client-facing agent work should begin read-only and task-bounded. Prompts are not security controls; consequential writes need technical scope limits, attribution, external activity evidence, and selective—not constant—human approval (The Agent Access Model, WriteGuard: fine-grained controls for MCP Servers).
  10. Charge from the first engagement and let closed deals justify each price increase; builder embarrassment about charging is the default failure mode, and a solo seller went from about $29 to $349 for the same product over four years while still calling it underpriced (Charge From Day One). Plan-first AI-assisted building doubles as a client discovery script (Plan Before Build (AI-assisted development)). Viral attention plays like Outbid.lol are luck, not a path, but their mechanism scales down to public local leaderboards and before/after scores that give participants a stake in promoting the venue (Launch Attention Loop).
  11. The Port Huron market has at least one local full-service web agency (Thomas Publishing House) and one Metro Detroit agency already selling an AI-visibility offer into local search (AlbaTech Services); the Chamber and EDA are the established convening channels, and the EDA has fresh small-business grant money and a new CEO as of September 2026 (Blue Water Area Chamber of Commerce, Economic Development Alliance of St. Clair County (EDA)).
  12. Published 2026 pricing puts a solo provider’s small automation builds at $500 to $5,000 with $150 to $800 per month retainers, and an established custom-AI shop’s floor at $7,500; survey data says the buyer objection to plan for is trust and error risk rather than cost (AI Services Pricing Benchmarks, SMB AI Adoption).

Open Questions

  • What service offering (and packaging) gets a solo AI-automation consultancy its first 3 paying clients fastest? Working answer (2026-08-12, unvalidated): the proof-business funnel — see Pivot brainstorm: the proof-business funnel.
  • How should services be priced for Port Huron / Blue Water Area SMBs? (Source’s $3k–10k audit pricing is B2B-SaaS-flavored — does it hold for small-town SMBs?) Partial direction (2026-09-03): price from the first engagement and raise in steps as receipts accumulate; early prices are almost always too low — see Charge From Day One. Dollar bands added 2026-09-04: paid discovery $500 to $2,500, small automation builds $500 to $5,000 (solo) or $7,500+ (agency), retainers $150 to $3,000 per month, Michigan website builds from $3,500 — see AI Services Pricing Benchmarks. Local willingness to pay still untested.
  • Which local industries (dental, trades, restaurants, manufacturing?) have the highest automation pain and willingness to pay? Partial (2026-09-04): surveys say 77 percent of US SMBs use AI but about 12 percent pay, payers are growth-focused and digital-first, entry use cases are admin, intake, and scheduling, and objections are trust-shaped, not price-shaped (SMB AI Adoption). County data says manufacturing and logistics dominate employment, but those firms are mid-sized (Economic Development Alliance of St. Clair County (EDA)). Vertical-level pain data still missing.
  • What’s the most effective local client-acquisition channel: referrals, chamber networking, cold outreach, or speculative rebuilds?
  • Are Cloudflare’s Pay-per-Crawl / x402 claims accurate and actually shipping? Answered 2026-08-12: yes, with maturity caveats — see Cloudflare.
  • When do Pay per Crawl, the Monetization Gateway, and Cloudflare Wallets reach GA, and does non-stablecoin settlement ship? (Wallets is announced but “coming soon” as of 2026-08-04 — see Announcing Cloudflare Wallets: The programmable wallet for the agentic Internet.)
  • What does the Blue Water Area Chamber charge for membership, and what do local agencies such as Thomas Publishing House actually charge? (Both have public pages not yet retrieved.)
  • What are St. Clair County’s establishment counts by sector? (Census County Business Patterns was blocked on 2026-09-04; the Economic Development Alliance of St. Clair County (EDA) employer table is the current proxy.)