Video Summary

How I Get Unlimited Leads Using Claude Code (For Cold Email)

Taylor Haren

Main takeaways
01

They built a proprietary lead-generation system using Claude Code that can process 272,000 rows per second (1M leads in ~5s).

02

Clay hit row limits and slow deletions, so the team migrated to a Claude Code + GitHub + Railway stack for scale and cost savings.

03

AI enrichment finds contacts across web sources (not just Google Maps) and achieves >95% valid-email rates.

04

Operational costs are low: Claude Code ≈ $200/month, overall stack about $1.9k–$2k/month; enrichment can cost ~$0.002 for three leads.

05

A private lead database (~50M leads) and automated refill keep campaigns supplied without constant data purchases.

Key moments
Questions answered

How fast can their Claude Code system process leads?

The custom system processes about 272,000 rows per second — roughly 1 million leads in 5 seconds.

Why did they stop relying on Clay?

Clay imposed row limits (50k rows/table, 12.5M workspace) and slow deletion times, which didn't scale for millions of leads or high-frequency processing.

What infrastructure do they use to host code and run workers?

They store code on GitHub and deploy processing workers on Railway, with lead data in Postgres (migrating toward Convex).

What are the running costs mentioned for the stack?

Claude Code is about $200/month; their total processing stack runs around $1.9k–$2k per month (Railway and workers included).

How reliable is the contact data after AI enrichment?

AI enrichment yields over 95% of contacts with a valid email and can find alternate valid emails when initial results fail.

Building a Custom Lead Generation Machine 00:25

"We built a custom lead generation machine that can process 272,000 leads per second."

  • The speaker discusses the creation of an efficient lead generation system using Claude Code, achieving an impressive capacity of processing 272,000 leads every second.

  • This capability dramatically reduces lead processing time, allowing the team to acquire 1 million leads in just 5 seconds.

  • The necessity for building their own system arose from limitations faced with existing tools, particularly Clay, which could not scale to the required level of performance as they approached 17.3 million platform accesses per week.

Transitioning from Clay to Claude Code 02:17

"For us, Clay limits you at 50,000 rows per table."

  • The video addresses the constraints imposed by Clay, which restricted users to 50,000 rows per table and a total of 12.5 million rows in the workspace.

  • The inefficiencies in deleting tables, taking days to complete, prompted the team to seek alternatives, leading them to Claude Code.

  • After experimenting with different platforms, switching to Claude Code proved to be a cost-effective and efficient solution, with expenses of only $200 per month.

Using GitHub and Railway for System Management 02:57

"All of our code sits inside GitHub."

  • The speaker explains that their entire code base is hosted on GitHub, which acts as a cloud storage system, ensuring everything remains organized and backed up.

  • They deploy workers on Railway, an innovative hosting platform that allows for efficient processing of code, akin to having functional "robots" that manage lead processing.

  • This setup has enabled the team to process approximately 12 million tasks per hour, showcasing the system's high efficiency.

Advantages of Their Custom System over Clay 06:15

"With our system, we can process 272,000 rows per second."

  • The custom system significantly outperforms Clay in terms of speed and efficiency, processing 1 million leads in just 5 seconds compared to 27 hours with Clay.

  • When campaigns need adjustments, the team can quickly pivot and reload new lists in real-time, a capability that most other tools lack.

  • The speaker highlights the stark contrast in operational efficiency, indicating that the custom system aligns perfectly with the demands of their cold email strategy.

AI Enrichment for Lead Generation 07:47

"We run an AI enrichment to find those contacts."

  • The process involves using AI to enrich data derived from Google Maps, scraping public databases to locate contacts at various companies.

  • This AI can determine critical attributes about a company, such as whether it is a multi-location medical practice and how long it has been in business.

  • This enrichment allows for a refined selection of leads based on Ideal Customer Profile (ICP) criteria before initiating outreach.

Cost Efficiency of Lead Acquisition 08:19

"Typically for each company, we want to find three leads, and the entire enrichment process costs a fifth of a penny."

  • The speaker emphasizes the low cost of acquiring three qualified leads through this AI enrichment process, amounting to just $0.002.

  • This shockingly low cost presents an efficient way to enhance lead generation without substantial financial outlay.

Enhancements to Data Sourcing 08:37

"We can use AI to find anybody who's on the internet at all."

  • The AI mechanism does not solely depend on Google Maps; it expands its search to capture contacts from across the web.

  • When using a vendor like AR Arc, if the initial search does not yield leads, the AI will explore other sources to find additional contacts, increasing the acquisition of qualified leads.

Comprehensive Lead Verification 09:59

"We’re able to get above 95% of contacts with a valid email."

  • The AI search provides significant enhancements in validity, allowing for over 95% of contacts to have verified emails compared to standard sources, which might yield only 30% reliability.

  • There’s an added layer of verification where if a valid email isn’t found for a lead, the system will attempt to identify alternative valid emails associated with that individual.

Targeted Advertising Insights 10:17

"We can scrape libraries at scale."

  • By scraping ad libraries from Google and LinkedIn, the system can identify companies running ads, suggesting they are focused on growth.

  • This creates a dataset of highly qualified leads that are more likely to be receptive to outreach, as they have dedicated resources toward client acquisition.

Executive Summary System for Campaign Analysis 11:01

"It gives us a daily report on everything that we should be thinking about."

  • The system automates the analysis of email campaigns, providing the team with daily updates on performance metrics such as lead-to-meeting ratios.

  • The goal is to refine email copy and strategies based on data-derived insights, helping improve the efficacy of future campaigns.

Private Lead Database Management 12:54

"We have our own private lead database with almost 50 million leads inside of it now."

  • The company maintains a comprehensive database of leads, allowing for minimal need to purchase new data since they can recycle and reassess existing leads.

  • This database provides insights into data vendor performance, enabling the team to optimize lead sourcing based on industry-specific efficacy.

Cost-Effective Autopilot for Lead Management 14:10

"Because of that automatic refill, we also know that clients will never run out of leads."

  • The lead management system is designed to automatically replace old leads with fresh ones, ensuring clients consistently have valid leads without incurring extra costs.

  • This automation in lead management sustains an efficient workflow, drastically reducing the overhead associated with lead acquisition and maintenance.

Overcoming Bottlenecks in Lead Generation 14:54

"The biggest bottleneck of our business was the processing speed of row limits, which we have now drastically reduced to just a few days."

  • The speaker highlights that the major challenge they faced was the slow processing speed related to row limits when generating leads. This bottleneck previously required them to spend significant time on pre-work, but they have since developed solutions to speed up this process.

  • They mention that now leads can be prepared in just a few days, which greatly enhances their ability to launch campaigns quickly. For instance, they share that clients on their accelerator plan can send out 5 million emails within 1 to 3 weeks of signing a contract, a remarkable improvement in efficiency.

Leveraging Claude Code for Automation 15:26

"You can just ask Claude Code how to integrate systems, and it will guide you step by step."

  • The speaker explains that Claude Code serves as a valuable resource for automating and integrating various tools within their lead generation processes. By asking detailed questions, users can obtain actionable insights and guidance.

  • They emphasize the importance of being willing to learn and adapt, regardless of one’s coding expertise. The focus should be on understanding objectives rather than code itself.

Building Internal Tools vs. Outsourcing Solutions 16:07

"If you're running cold email at scale, you have two paths forward: build your own systems or leverage our established ones."

  • The speaker discusses two distinct paths for businesses looking to effectively manage cold email campaigns at scale. The first is to build a customized internal system using Claude Code, where businesses can address their most pressing challenges incrementally.

  • The alternative is to utilize the existing systems developed by the speaker's team, which can save time and effort for businesses focused on outcomes. They provide a call to action for viewers interested in leveraging their proven strategies, showcasing successful case studies.

Results Achieved Through Cold Email Strategies 16:40

"We've helped companies like RB2B scale to $4 million in revenue in just four months, with 42% of that revenue coming from our cold emails."

  • The speaker shares impressive results from using their cold email strategies, including a significant increase in annual reoccurring revenue for clients within a short timeframe. They credit a substantial portion of this success to the cold emails they initiated.

  • Specific metrics are highlighted, such as generating $4.3 million in annual pipeline for Fixer AI. The potential for massive growth is illustrated through estimates of what could be achieved if outreach efforts covered the total addressable market every two months.