AI Sales Agent Developer — Outbound Lead Generation + CRM Data Analysis for B2B SMB

We’re a US-based B2B company looking to build an AI-powered sales agent that handles outbound prospecting, personalized outreach, and lead data analysis — so our sales team spends time closing, not researching. This is a hands-on build role, not a consulting engagement. We want working software, not a strategy deck. What We Need Built 1. Outbound Lead Engine • Pulls prospect data from sources like Apollo, LinkedIn Sales Navigator, or Hunter • Enriches contacts with company size, tech stack, job titles, buying signals • Scores and prioritizes leads based on ICP criteria we define 2. Personalized Outreach Automation • Drafts first-touch cold emails with context-aware personalization (not mail-merge) • Manages follow-up sequences across email (and optionally LinkedIn) • Respects sending limits and domain health (no burning our domain) 3. Lead Analysis Dashboard • Tracks open rates, reply rates, positive vs negative responses • Surfaces patterns: which ICP segments respond, which messages convert • Flags leads that show buying intent signals for human follow-up 4. CRM Integration • Pushes qualified leads and conversation history into HubSpot or similar • Updates lead status automatically based on replies Tech Stack We’re Open To • N8N, Make, or custom Python for orchestration • OpenAI / Anthropic APIs for message drafting and classification • Apollo, Clay, or equivalent for enrichment • Smartlead, Instantly, or similar for sending infrastructure Must Have • Delivered at least one working outbound AI agent, not just prototypes • Can show actual reply rates or pipeline data from a past build • Understands email deliverability basics (SPF, DKIM, warm-up, sending volume) • Clear documentation and handoff — we need to own and maintain this Nice to Have • Experience with US SMB outbound specifically (not just enterprise SaaS) • Knowledge of Clay + Apollo combo workflows • Prior work with CRM auto-enrichment Project Scope • Fixed-price milestone-based engagement preferred • Start with Phase 1 (lead engine + outreach automation), expand if it works • Budget range: open to proposals, but include past results not just past projects To Apply Share one past AI agent build with the outcome (reply rates, leads generated, revenue influenced). If you can’t share numbers, share the architecture and explain what worked and what didn’t. We don’t need polished pitches — we need people who’ve done this and know where it breaks.