Automating the Entire Consulting Sales Pipeline with AI Agents

Problem

Traditional consulting sales pipelines are labor-intensive and reactive. Prospecting for new opportunities requires manual job board scanning across multiple platforms, individual assessment of lead quality, manual CRM entry, and repetitive follow-up coordination.

The baseline challenge: prospecting consumed 10+ hours per week of senior consultant time across job board monitoring, lead qualification, proposal drafting, and pipeline management. This created lost market intelligence as opportunities were missed, and opportunity cost of expert time diverted from billable work.

Approach

Rather than optimize individual steps, we automated the entire pipeline end-to-end using coordinated AI agents operating on three core layers:

Market Intelligence Layer

A Job Hunter agent scrapes 19 specialized job board sources across three categories (AI/ML projects, full-stack development, and consulting engagements). This agent runs hourly via Groq's llama-3.3-70b inference API, leveraging the free tier for cost-effective real-time alerts.

Opportunity Processing Layer

Raw postings flow into automated analysis that evaluates market fit, client profile quality, and engagement likelihood. Results are categorized, scored, and persisted to Google Sheets via service account integration. The CRM becomes a living, continuously updated database.

Outreach Execution Layer

A second agent layer generates personalized cold outreach emails by synthesizing insights from nightly intelligence briefs. Follow-up scheduling is automated based on engagement patterns and pipeline stage.

Architecture

Three coordinated but independent agent workflows:

• Job Scraping Agent: Monitors 19 sources hourly, normalizes postings into a standard schema, filters for relevance, outputs to Google Sheets via the Sheets API
• Analysis & Scoring Agent: Evaluates each posting for client fit, project scope, and engagement likelihood with categorization (AI/ML, full-stack, consulting) and confidence scores
• Outreach & CRM Agent: Consumes analysis outputs and nightly intelligence briefs to draft personalized outreach, create CRM records with stage tracking, and schedule follow-ups automatically

All agents leverage lightweight LLM inference (Groq's free tier) for cost efficiency. The system is deployed on Railway with a web UI, streaming responses, and health monitoring.

Results

The automated pipeline transformed consulting sales from a time-consuming manual process into a continuous, low-touch system. Each night, the intelligence brief identifies emerging patterns that inform the next day's outreach positioning -- creating a feedback loop where market insights continuously improve proposal relevance.

Key Metrics