AI Agent for Lead Management

I am looking for an automation/AI workflow developer to help build a lead enrichment and qualification system for an acquisition/origination platform. Current workflow: We export raw company lead data from ReferenceUSA / Data Axle. The export contains many fields we do not need. We need to clean the file, extract only the required fields, enrich each company online, qualify the lead, and automatically send qualified records into Airtable. What I need built: 1. Data Cleaning / Extraction * Take raw Data Axle Excel/CSV exports * Remove unnecessary fields * Map useful fields into a clean standardized format * Create a clean lead input file for enrichment 2. Web Enrichment For each company, the system should search online and find: * Correct company website * Owner / Founder / President / CEO * Decision-maker email if available * Company LinkedIn page * Owner/executive LinkedIn profile if available * Services offered * Location verification * Private/family-owned signals * Public company, franchise, or PE-backed red flags * Source URLs for evidence 3. Qualification Logic The system should classify each company as: * Pass * Needs Review * Reject The system should also produce: * Fit score * Confidence score * Reason for qualification * Missing data * Evidence URLs 4. Airtable Integration * Create or update records in Airtable * Avoid duplicates * Only send qualified or “Needs Review” leads into the main Airtable table * Keep rejected leads separate or in a rejected log Preferred tools: * n8n, Make, or Zapier * Airtable API * Python * OpenAI / Claude API * Perplexity API or other web research tools * Browser automation tools such as Playwright, HyperAgent, Browserbase, or similar Ideal candidate: * Has built lead enrichment workflows before * Understands Airtable bases, APIs, and deduplication * Can work with messy Excel/CSV files * Can create structured JSON outputs * Can build reliable automations with error handling * Understands that accuracy and evidence matter more than volume Please include: * Examples of similar workflows you have built * Your recommended tech stack * Whether you would use n8n, Make, Python, or another approach * Estimated timeline for an MVP * Estimated cost for building the first version The first version should process a test batch of 100–250 companies before scaling.