the gtm engineer AI Voice Agents & Workflows that Convert with Manthan Patel, founder at Lead Gen Man
Manthan Patel began working with YC founders two years ago, where he first started using AI agents and LLM workflows. When AI agents gained popularity, Manthan already had hands-on experience, so he started recording and sharing what he was building. In January 2025, he began posting content on LinkedIn and grew his following from zero to 100K in just six months. Today, Manthan runs both an AI automation agency, and a lead gen agency. His agencies offer low ticket courses that over 50,000 people have taken, while also offering white glove agent-building and implementation services.
Subscribe for weekly updates on top GTM Engineering content, open roles & more
In this podcast, we discuss:
* How frequency of posting is one of the last remaining differentiators on LinkedIn
* How to repurpose content across LinkedIn, Instagram, and TikTok in platform-specific formats to maximize distribution
* Building effective AI voice agents for outbound and inbound calls
* Automated inbound form enrichment workflows that qualify leads to prevent SDRs wasting time on unqualified calls
* Building end-to-end prospecting workflows using Claude MCP
* When to use Clay vs. n8n
* Self-hosting LLM infrastructure to reduce API costs, maintain control of data, and meet compliance requirements
Episode highlights:
* Manthan grew his LinkedIn following from zero to 100K in six months by posting lead magnets that encouraged comments to amplify their reach. He also posts three to four times per day across different global time zones to maximize visibility and create faster feedback loops on what content resonates with his audience.
* Manthan repurposes the same content across LinkedIn, Instagram, and TikTok by adapting it to each platform’s format. For instance, he’ll convert LinkedIn carousel posts into short-form videos for Instagram and TikTok. By being present on multiple platforms he’s able to reach his audiences where they actually consume content.
* Manthan ran an automated cold-calling campaign for a vending machine company by scraping local business data, and personalizing AI calls with shop names and addresses. With this personalization along with multiple call attempts, and disclosing upfront that the call was coming from an AI agent, Manthan generated 80 demos across 10,000 leads in 1 month.
* Manthan built an inbound form enrichment workflow where prospects submit only their name and email, then an AI agent enriches both personal and company data, feeds it to a second AI that evaluates ICP fit, and only books demos with qualified leads. This prevents SDRs from wasting time on calls with unqualified prospects who fill out forms.
* Manthan uses Claude with MCP servers to do prompt prospecting, where he describes research tasks in natural language. Prompted with these research tasks, Claude connects to data APIs like Lusha to find prospect information, and once found, Manthan has Claude add these leads directly to HubSpot. This eliminates the need to manually build workflows or switch between multiple tools for prospecting or tracking.
* An emerging trend Manthan sees is self-hosting LLM infrastructure. Compliance-focused clients run models locally on hardware like a Mac Mini to avoid sending their data to big models like OpenAI or Anthropic. This approach allows these orgs to take advantage of AI while preventing data exposure that could trigger compliance audits and license loss for regulated industries, while also reducing long-term API costs for high volume AI usage.
Where to find Manthan:
Transcript details:
(00:00) Intro and background
(02:27) Manthan’s personal branding and agency structure
(03:44) Growing his LinkedIn from zero to 100K in six months
(05:54) Manthan’s early client work
(08:48) Learning LinkedIn strategy
(12:30) Building social media presence across multiple platforms
(15:12) AI voice agents at scale to drive 80 demos from 10,000 AI cold calls
(18:59) Backtesting prompts and handling edge cases
(20:10) Manthan’s AI agent tech stack
(24:28) Prospects don’t care when disclosing AI upfront
(28:12) Inbound AI agents for 24/7 support with conversation history
(30:38) Nailing AI call agent prompting
(31:58) When to use Clay versus n8n
(33:44) Self-hosted n8n for compliance-driven enterprise clients
(35:00) Manthan’s favorite workflows
(37:48) What 50K people have learned from Manthan’s courses
(40:16) Using Claude with MCP servers for prompt prospecting
(42:35) Local LLM infrastructure for compliance and cost
(44:39) How to get started with workflows, agents, and MCPs
(45:33) Favorite underrated tool, growth hack, and conclusion
For inquiries about sponsoring the podcast and to recommend any guests, email noah@thegtmengineer.ai
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit thegtmengineer.substack.com