← Back to Daily Builds

Daily build note · June 22, 2026

Agency AI Visibility Pilot Pack

A deterministic packaging layer over the Local AI Visibility Watchboard that turns generated local AI visibility packets into a productized agency consulting offer: one-pager, pilot offer...

Consulting Offers / Productized Services Build note published Public demo coming soon

What shipped

Built agency_ai_visibility_pilot_pack, a standard-library Python CLI that generates an agency sales pack from a local Local AI Visibility Watchboard workspace.

Commands:

  • python -m agency_ai_visibility_pilot_pack sample --path /tmp/agency-ai-visibility-pilot-pack-demo
  • python -m agency_ai_visibility_pilot_pack pilot-pack --path /tmp/agency-ai-visibility-pilot-pack-demo
  • python -m agency_ai_visibility_pilot_pack validate-pilot-pack --path /tmp/agency-ai-visibility-pilot-pack-demo

Generated outputs include the agency one-pager, $300 pilot offer, intake checklist, prospect shortlist CSV, outreach sequence, discovery script, pricing menu, sample SOW, handoff tracker CSV, and pilot-pack.json.

Architecture

  • Standard library only: argparse, json, csv, pathlib, datetime, and unittest.
  • Local deterministic file generator instead of a web UI or hosted service.
  • watchboard/watchboard.json is treated as the primary parent contract.
  • The sample command mirrors the parent demo businesses and writes a minimal compatible watchboard workspace so the product runs without importing or modifying the parent build.
  • Validation checks structure, demo coverage, pricing, and compliance language instead of requiring network calls or live data.

Trimmed scope

Intentionally excluded:

  • Web UI
  • Database
  • Auth
  • Billing
  • CRM sync
  • Email sending
  • Browser automation
  • Live ranking checks
  • Google Business Profile API
  • Any OpenAI, Claude, Anthropic, OpenRouter, Gemini, or other model/provider call
  • Automated review posting or profile publishing

Limitations

  • The sample watchboard is a compact local contract, not a full re-run of the parent watchboard scorer.
  • Templates are deterministic and practical, but not brand-customized.
  • Validation checks for required language and obvious forbidden promises; it is not a legal or platform-policy review.
  • Source freshness is documented as a human workflow requirement, not checked against live sources.

Suggested next steps

  • Add optional template overrides for agency branding and target verticals.
  • Add a thin import adapter for real parent watchboard workspaces if the contract expands.
  • Add a redline-style review report that highlights missing client evidence before pack generation.
  • Add more seeded verticals after the first paid pilot proves which artifacts are actually used.