The Problem

There are 33 million small businesses in the United States. Most depend on Google to find new customers. Staying visible in local search requires constant upkeep — weekly posts, accurate information, review responses, and profile optimization. Most small business owners have no time for this. A marketing agency costs $500-$2,000 per month. Most small businesses cannot justify that cost.

The result is a silent, compounding revenue leak. A dental practice with a stale Google listing loses potential new patients every day to a competitor who posts more consistently. The gap between what a business could earn and what it actually earns from local search is real, measurable, and fixable — but only if someone is watching and acting continuously.

Nobody is.

What GBPilot Does

GBPilot is an AI-native Google Business Profile management service. For $79 per month per location, GBPilot's AI agents audit a business's listing, identify specific revenue-losing gaps, generate and publish weekly content, monitor daily ranking position, and alert the owner when action is needed — all without human intervention.

How AI Runs the Business

GBPilot operates through four specialized agents running continuously in production on Google Cloud Vertex AI with Gemini 2.5 Flash:

Audit Agent — Queries Google Maps Places API, scores listings against 40+ local SEO ranking factors, and outputs a dollar-denominated revenue loss estimate. Every decision is logged to structured JSONL telemetry with timestamp, input signal, and output.

Content Agent — Drafts Google Business Profile posts, Q&A responses, and profile updates in the business's voice, generated by Gemini based on business category and competitive landscape.

Outreach Agent — Scrapes target businesses from Google Maps, generates personalized audit reports, and sends AI-authored emails referencing the specific dollar amount each business is losing. Contacted 36 real US businesses across Ohio, Indiana, Kentucky, and Tennessee autonomously.

Monitoring Agent — Tracks daily ranking position, detects drops, and triggers re-audits when competitors gain ground — running continuously without human scheduling.

What Humans Do

The founder handles three things: strategic product decisions, pricing, and customer support escalations. Everything operational is handled by AI agents.

What Was Built

  • Live product at gbpilot.onrender.com
  • Full audit-to-payment flow working end to end
  • Paystack live payment integration active
  • 36 real US businesses contacted by AI outreach pipeline
  • 500+ Gemini 2.5 Flash API calls made in production
  • Complete agent decision telemetry logs available for review

Challenges

The biggest challenge was infrastructure access — Nigerian payment cards trigger a $30 prepayment hold on new Google Cloud billing accounts, blocking Vertex AI access entirely. The solution was authenticating via personal OAuth refresh tokens on a colleague's GCP project, which required days of troubleshooting across service account permissions, org policies, and authentication scope issues. Every problem was solved without any paid tools or outside contractors.

What I Learned

Building an AI-native business in 90 days as a solo developer from Nigeria taught me that the hardest problems are not technical — they are distribution. The product works. Getting it in front of the right people across timezone, trust, and channel barriers is the real challenge that AI has not yet solved.

Built With

  • fastapi
  • gemini-2.5-flash
  • google-cloud-vertex-ai
  • google-maps-places-api
  • google-oauth2
  • paystack
  • python
  • render
Share this project:

Updates