AI Sales Lead Generator

We were inspired to build the AI Sales Lead Generator after speaking with startup founders and interns who spent up to eight hours a day manually updating lead sheets, sending cold emails, and tracking responses. While enterprise CRMs such as Salesforce and Copilot Sales are incredibly powerful, they often feel bloated and unintuitive for smaller teams who need agility, personalization, and far more reach. We wanted to create something different; an intelligent, lightweight, and fully automated system that would empower startups and niche B2B sales teams to compete with larger organizations. Our vision was to design a tool that blended AI-driven personalization with automation, focusing on LinkedIn and niche market outreach rather than broad, generic lead generation.

The AI Sales Lead Generator automates the entire B2B sales workflow from start to finish. It discovers potential leads from Shopify and LinkedIn, enriches their company data through external APIs, and stores everything in DynamoDB for structured access. Using Amazon Bedrock’s Nova model, it generates personalized, context-aware outreach emails tailored to each company and automatically sends them through Amazon SES. When a reply comes in, it is received by SES, stored in S3, and triggers a Lambda function that classifies the sentiment as Warm, Cold, or Neutral before updating DynamoDB in real time. A Streamlit dashboard visualizes all this data, allowing users to monitor campaign performance and engagement metrics instantly. The entire system is serverless, event-driven, and self-scaling, requiring no manual supervision while maintaining full transparency and control.

We built this platform on a serverless AWS stack, integrating API Gateway, Lambda, DynamoDB, SES, S3, and Bedrock to handle the full data flow. Each Lambda function is dedicated to a specific task: lead discovery, enrichment, email generation, sending, and reply classification. By connecting these components through event triggers, we achieved an end-to-end automated workflow. The Streamlit dashboard communicates with API Gateway to fetch live campaign insights, while DynamoDB ensures consistent, low-latency access to leads and replies. This architecture makes the system scalable, cost-efficient, and capable of handling thousands of leads simultaneously without any infrastructure management.

Of course, building such a complex, interconnected system came with challenges. Ensuring reliable communication between SES, S3, and Lambda required extensive debugging and event mapping, especially for inbound email processing. Optimizing Bedrock’s AI output to maintain a human-like yet professional tone was another hurdle; striking a balance between personalization and consistency was key. We also had to carefully design DynamoDB’s schema to accommodate different types of data, leads, replies, and campaign stats, while maintaining query performance and scalability. These challenges taught us the importance of architecture-first thinking, where every design decision has downstream effects on reliability and user experience.

We’re proud that our product is working and has the ability to be more fleshed out and scalable. It achieves tangible results: 10x faster lead processing, 90% fewer manual tasks, 5x greater personalization, and 3–5x higher response rates. More importantly, it proves that AI and automation can be powerful equalizers, giving smaller teams the same efficiency and precision as enterprise tools without the complexity or cost. Watching the workflow run entirely on its own, from lead discovery to reply classification, was a major milestone for us.

Through this project, we learned how to build intelligent systems that scale efficiently and integrate AI meaningfully into automation. We gained deep experience in serverless design, event-driven programming, AI model integration, and full-stack automation, as well as a better understanding of how to design products for real-world use cases rather than just technical demonstrations.

Next, we plan to take this project even further. Our roadmap includes LinkedIn integration, CRM syncing, and a multi-user SaaS dashboard that will allow entire sales teams to collaborate, track campaigns, and automate follow-ups. We’re also exploring Bedrock Agents for self-improving AI-driven lead scoring and message sequencing. Ultimately, our goal is to make the AI Sales Lead Generator a plug-and-play solution that any startup, freelancer, or B2B team can use to transform their outreach, freeing people from tedious manual work and letting AI handle the repetitive grind.

Built With

• bedrock
• dynamodb
• lambda
• s3
• ses