Local Business Lead Discovery and Enrichment Agent

Description

A weekly agent that finds local businesses with weak or missing web presence, and scores them as sales leads for web agencies, freelance developers, and digital consultants.

The agent searches local business directories in your target area. It scrapes each business website with Firecrawl, evaluates quality signals, scores opportunity as High / Medium / Low, deduplicates against a database so you never process the same business twice, and sends you a clean weekly email with fresh leads ready to act on.

What problem is it solving?

Finding clients who genuinely need your help is the hardest part of running a web agency or freelance business. Cold outreach to random businesses wastes time. This workflow automates prospecting: it identifies businesses that have strong reviews but a weak digital presence.

The deduplication layer means every lead in your inbox is genuinely new. The scoring system means you can focus on High-opportunity leads first and work down as needed.

What this workflow does

Triggers on a weekly schedule (Mondays at 9:00 AM)
Randomises the search seed each run to explore different categories and areas, reducing repetition across weeks
AI agent searches local business directories for target business types in the configured area
Firecrawl scrapes each business website to evaluate quality signals:
Mobile-friendliness
Outdated design indicators
Missing key pages (contact, services, booking)
Tech signals (old frameworks, no HTTPS)
Freshness (last updated, stale content)
Scores each lead as High / Medium / Low opportunity, with a written reason for why it's a lead
Deduplicates against the marco_leads Postgres table — businesses already in the database are skipped
Saves new leads to the database with full enrichment data
Sends a Gmail digest with the week's new leads, formatted for quick review

Setup

Create the database table — run the SQL provided in the workflow's sticky note to create the marco_leads table in Supabase (or any Postgres database)
Connect Postgres credentials — point them at your Supabase project
Connect Firecrawl credentials — used for directory and website scraping
Connect OpenAI credentials — used by the AI agent for discovery and scoring
Configure Gmail — set your Gmail address in the Gmail node as the sender and recipient
Adjust the target parameters — update the AI agent's instructions with your preferred city, radius, and business types
Activate the workflow — it will run automatically every Monday at 9:00 AM

How to customize this workflow to your needs

Change the target location** — update the city and radius in the AI agent prompt; the workflow works for any geography
Change business types** — swap dentists, salons, gyms, and restaurants for any local business category relevant to your market (e.g. lawyers, accountants, florists)
Adjust lead scoring criteria** — modify the agent prompt to emphasise the signals that matter most to your outreach strategy
Change the output** — replace the Gmail node with a Slack message, a Notion database entry, or a CRM webhook to slot leads directly into your sales pipeline
Change the schedule** — run it daily for higher volume, or bi-weekly if you have a smaller target area
Add enrichment** — extend the workflow with a phone/email lookup step after scoring to enrich High-opportunity leads with contact details before they hit your inbox
Filter by opportunity score** — add a filter node before the Gmail step to only send High-opportunity leads, saving review time when volume is high