LOCAL BUSINESS AdRails 0.8.0

Leads from Google and AI, run by your own Claude

AdRails 0.6.0 adds 12 tools for local service businesses: Google Local Services Ads leads, reports, junk-lead feedback and verification readiness; a lead book for every channel; a generic 90-day growth plan that each business's Claude customises; LocalBusiness schema, name/address/phone checks and service-area page briefs for AI answers; and a reviews engine. It runs on your computer with your own Google sign-in.

Last reviewed 2026-11-06 with AdRails 0.8.0

14 TOOLS · 5 PLAYBOOKS · 4 READY-MADE JOBS

What can a small business's Claude do with AdRails?

Every tool returns plain data that Claude explains in plain English. The four prompts (start, morning, weekly, monthly) chain the tools into the routine, so the owner never has to know what to ask.

Google Local Services Ads

  • adsops_lsa_leads: leads with call and message history, answered or missed, first-response minutes, junk flags. Each lead syncs into the lead book.
  • adsops_lsa_report: spend, cost per charged lead against the most you can afford, credits, booking rate, missed calls, leads by hour and weekday.
  • adsops_lsa_dispute: suggested ratings (junk = dissatisfied with Google's reason; booked = very satisfied). Submitted through ProvideLeadFeedback only with commit=true.
  • adsops_lsa_readiness: license, insurance, background check and registration status, expiry warnings, and the profile signals Google ranks on.

Leads and the plan

  • adsops_local_profile: the questions to ask the owner, saved once, plus the most a lead can cost before it loses money.
  • adsops_local_plan: the generic 90-day plan, customised from the profile, with progress tracking.
  • adsops_leads and adsops_leads_due: one lead book for every channel, stages from new to won, follow-ups, pipeline by source, CSV export, and won jobs ready for offline conversion upload.

Found by Google and AI assistants

  • adsops_local_schema: LocalBusiness JSON-LD with the right type for the trade, service areas, services, hours and license. It never adds self-serving review stars.
  • adsops_local_nap: name, address and phone consistency across listing pages, plus the citation checklist (Apple Business Connect and Bing Places feed Siri, ChatGPT and Copilot).
  • adsops_local_area_page: a brief for each service × area page, built around the questions people ask AI assistants.
  • adsops_reviews: review link, request drafts, reply drafts that flag legal and safety reviews, and review velocity against competitors.

Live status: every LSA query (leads, call and message history, campaign spend, verification) runs against the live Google Ads API v24, verified 6 November 2026. That first live run caught a field Google rejects, and a contract check now pins every field to Google's own list. The suite passes 84 integration checks through the real server, and the Google sign-in flow is tested over a real local socket.

PLAYBOOKS

Where does the strategy come from?

Tools audit and draft; strategy decides what to do first. AdRails ships five playbooks written for your Claude to execute, distilled from weeks of work on real businesses: AEO, local leads, Google Business Profile, Google Ads and ChatGPT Ads. Every task in the 90-day plan links the section behind it.

AEO: getting named by AI assistants

  • A 7-day fast path for a business with almost no web presence.
  • The visibility panel: 8 to 15 locked, unbranded, situation-shaped questions run in clean sessions. List-style questions under-report what AI actually sends you.
  • Pages engines quote: a 45 to 70 word answer first, real numbers, a few strong service-area pages instead of a thin wave.
  • Myths, labelled: every claim is marked tested or early evidence. No one can promise a citation, and the playbook says so.

Local leads: booking the jobs

  • Speed to lead twice a day, and every public phone number tested.
  • Follow-up cadence around day 0, 3, 7, 14, 28 and 45, 72 hours apart, and any reply pauses it.
  • LSA, Google Business Profile and reviews in the order that pays, with the checks that catch the expensive mistakes.
  • Business Profile to 100: adsops_gbp scores the profile, ranks the fixes and checks the description against Google's rules.
  • Before and after: adsops_results compares the month before AdRails with now, and never invents a number it does not have.
  • Read them with adsops_playbook. AdRails tells your Claude when a new release is out (adsops_update_check), so improvements reach you.
RAILS

What can't it do?

  • No messages to customers. Review requests and replies are drafts. Google's API offers no LSA consumer messaging, and AdRails doesn't pretend it does.
  • No surprise spend. Search campaigns change only through a committed plan with caps you set. LSA budgets live in Google's dashboard.
  • Your data stays yours. The lead book, profile and Google sign-in are files on your computer (chmod 600). Customer names and phones never enter the audit log.
FAQ

Straight answers.

Can Claude manage Google Local Services Ads?

Partly, and AdRails is exact about which part. Through the Google Ads API it reads LSA leads, their calls and messages, charge and credit status, campaign spend and verification status, and it submits lead feedback, which is how Google decides credits for junk leads. Google does not offer LSA budgets, bids, hours or pausing through the API, so those stay in the LSA dashboard. Every LSA query runs against the live Google Ads API v24 (verified 6 November 2026).

How does AdRails decide a lead is junk?

It reads the lead's messages and calls and flags sales pitches (SEO, listing and financing offers), duplicates from the same phone within 7 days, very short calls, out-of-area places and jobs you don't do. Booked leads are never flagged. Each flag maps to Google's own feedback reason, and nothing is submitted until the owner approves.

What is in the 90-day plan?

Four stages: foundations (Google Business Profile, consistent listings, schema, lead book, tracking), get leads now (Local Services Ads, a capped Search campaign if there is budget, reviews, a 5-minute response rule), get found by AI (service-area pages, FAQ answers, llms.txt, citations on Bing, Apple and Yelp), and optimise (lead feedback, cost per lead against the most you can afford, landing pages, offline conversions). It skips what you already have and drops ad tasks when the budget is zero.

Does it send review requests to my customers?

No. It writes the request and reply drafts and tells you the rules: ask every customer, never only the happy ones, no incentives, send within a day of the job. You or your phone sends them.

How do I start?

Paste one prompt into Claude Desktop: Set me up with AdRails for my business. Read https://adrails.dev/start/ and follow the steps for Claude. The start page lists every step Claude runs.