Audit My AI Agency Lead Generation System
You are a senior AI automation architect, outbound systems engineer, and lead qualification auditor.
I am building an AI-powered agency system that finds local business leads, qualifies them, drafts outreach, and eventually helps create website demos. Before I automate outreach, I need you to audit my top-of-funnel lead generation system for quality, safety, and revenue potential.
Important:Do not change code yet. Report only.Do not enable sending.Do not create Gmail drafts.Do not trigger demo builds.Do not post to Slack or Telegram unless I explicitly ask.Do not modify existing production workflows. Audit the system across two lead lanes:
Lane 1: Google Maps No-Website LeadsThese are businesses found through Google Maps or a similar local search source that appear to have reviews, proof of activity, and no owned website.
Audit for:
- How leads are collected.
- Whether the system confirms the business has no owned website.
- Whether review count and rating are used as demand signals.
- Whether the system deduplicates by business name, phone, email, domain, and address.
- Whether the system checks query relevance, so a search for “window cleaners” does not qualify generic house cleaners.
- Whether leads without email are routed correctly to phone review, enrichment, or manual review.
- Whether the system avoids forcing no-email leads into an email campaign.
- Whether bad categories are suppressed, such as directories, schools, government, nonprofits, spam, or irrelevant businesses.
- Whether every run logs source analytics, such as source query, source run ID, qualified count, suppressed count, and manual review count.
Recommended statuses:
- Raw_Scraped
- Needs_Enrichment
- Qualified_Phone_Review
- Manual_Review
- Suppressed
- Rejected
- Active Dialogue
- Draft_Created
- Contacted
Lane 2: Slow Website / PageSpeed LeadsThese are businesses that already have websites, but the sites may be slow, weak on mobile, outdated, or poorly optimized.
Audit for:
- How website leads are collected.
- Whether emails are valid and contactable.
- Whether PageSpeed or mobile performance evidence is collected.
- Whether technical issues are translated into business pain, such as lost calls, missed quote requests, or mobile visitors leaving.
- Whether the audited URL is the business’s owned website, not a directory, booking portal, article, association profile, or third-party platform.
- Whether the system blocks weak leads, including:
- Whether the system correctly suppresses fast sites.
- Whether PageSpeed timeouts route to Manual_Review instead of being labeled dead links.
- Whether API keys and secrets are redacted from logs.
- Whether qualified leads are routed to a manual approval queue before any draft or send action.
Minimum safety requirements:
- No autonomous sending.
- No broad Gmail draft mode.
- No batch drafts without manual review.
- No demo build unless the lead has clear intent or I manually approve it.
- No last_email_copy update unless draft creation succeeds.
- All sensitive API keys must be redacted in logs.
- Dry-run mode must exist before production mode.
Please deliver:
- Current system summary.
- Top 10 risks.
- Lead quality failure modes.
- Safety failure modes.
- Suggested statuses and routing.
- Suggested database fields or score_breakdown fields.
- Suggested feature flags.
- Recommended tests.
- Recommended dry-run commands.
- Recommended implementation order.
- A clear “do not proceed to outreach until these pass” checklist.
End with a concise roadmap for turning this into a safe lead-generation engine.