Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/vrashmanyu605-eng/Agentic_Sales-Markerting/llms.txt

Use this file to discover all available pages before exploring further.

Agentic Sales & Marketing is a multi-agent AI system built on LangGraph that automates the full B2B sales pipeline for IT services companies. You configure it once with your company profile, pricing, case studies, and a target industry — then the system discovers leads, researches each one, scores it against your ideal customer profile, analyzes the competitive landscape, generates personalized outreach and proposals, and writes everything to Google Sheets. The system is designed for IT services and AI automation companies looking to scale outbound sales without manual research overhead.

Key features

Lead discovery

Web search-powered discovery of target companies based on your industry, requirements, and client job description file.

Lead research

Deep research on each discovered company: website scraping, LinkedIn decision-maker lookup, and company intelligence gathering.

ICP matching

Score each lead against your configured ideal customer profile and filter out poor fits before investing time in outreach.

Competitor analysis

Map the competitive landscape for each deal and surface positioning strategies tailored to the lead’s existing vendors.

Outreach generation

Generate personalized cold emails, LinkedIn messages, follow-up sequences, and call pitches for every qualified lead.

Proposal generation

Produce structured IT services proposals with scope, implementation plans, and pricing drawn from your sales deck.

CRM update

Automatically write lead details, research summaries, outreach content, and next steps to your Google Sheet.

Agent pipeline

The workflow runs as a deterministic LangGraph pipeline. A supervisor agent orchestrates a queue of discovered leads, routing each one through the same sequential chain of specialized agents.
1

Lead discovery

The discovery agent is the entry point. It searches the web for companies matching your target_industry and client_requirements, then builds a queue of candidate leads stored in discovered_leads.
2

Supervisor routing

The supervisor inspects the lead queue. For each unprocessed lead it routes to lead_research_agent. When the queue is empty, it sends the graph to END.
3

Research and qualification

The lead research agent gathers intelligence on the active lead. The ICP matching agent then scores it against ideal_customer_profile and records the analysis in icp_analysis.
4

Competitive intelligence

The competitor analysis agent uses competitors_data and the lead’s profile to map competing vendors and identify differentiation opportunities.
5

Content generation

The outreach generation agent writes personalized messaging using the lead research, ICP score, and your sender_name. The proposal generation agent drafts a structured proposal drawing from sales_deck_text, pricing_data, and case_studies.
6

CRM update

The CRM update agent appends all outputs — research, analysis, outreach, and proposal — to your Google Sheet, then hands control back to the supervisor for the next lead.
The system uses LM Studio running locally with the gemma-3-4b-it model by default. You must have LM Studio installed and the model loaded before running main.py. The LLM client connects to http://localhost:1234/v1. See Configure the sales workflow state and LLM for details.

Build docs developers (and LLMs) love