Manual Prospecting to AI Agent for Lead Scoring: SmartCat’s SDR Transformation Story

60 minutes. That’s how long it took one SDR to research, score, and write an email for a single lead. Multiply that by dozens of leads a week, and it’s clear to see the scope of the challenge our sales team faced.

Despite the skill and effort of our SDRs, the process was too manual, too slow, and too dependent on personal judgment. Besides, sales development is no longer about sending more emails – it’s about sending the right ones.

That’s why built an internal AI prospector that turned those 60 minutes into just a few clicks. Instead of adding more people, we gave our existing team a smarter, more efficient AI-powered partner.

Here’s how we used automation to transform prospecting from guesswork to strategy. 

Problem

Our sales development representatives (SDRs) were spending their time, skills, and energy on leads that didn’t match our culture, values, goals, and vision. As a result of that mismatch, there were too many cold leads.

Also, our sales team was doing all of the prospecting manually, which came with its own set of issues:

• Time drain: 60+ minutes of research, writing, and prep for every single lead.
• Inconsistency: Each sales rep approached qualification differently, leading to uneven results.
• Low engagement: Outreach lacked personalization at scale, making it harder to stand out.
• Missed opportunities: Without an objective scoring system, good leads were sometimes overlooked

SDRs were casting their nets wide, but some of the catch kept slipping through the gaps.

Solution

To narrow down the aforementioned gaps, we decided to create a web-based AI scoring tool to analyze and score each lead based on the formula defined by the sales team.

It acts like a digital sales assistant, handling the repetitive work so reps can focus on conversations.

More specifically, this AI prospector tool automates and enhances the following areas of cold outbound:

Lead Scoring

Using GPT-4.1 with web browsing, we compare company information manually entered by our sales reps (company name, contact name, website) with a generated company description. 

A structured prompt evaluates fit across multiple criteria, such as industry match, product relevance, and company size, to come up with a lead score in a consistent format, taking the guesswork out of the equation.

Personalized Cold Email Generation

Once the right lead is identified, the next step is connection. A second GPT-based prompt takes over and crafts a personalized cold email in our brand’s voice. 

Each message feels written for the reader, not at them, focusing on their challenges and goals, while applying the best copywriting practices, such as:

• Interest-based CTAs: They invite the leads, not push the message onto them.
• Pattern interruptions:  Copy that breaks through inbox noise.
• “You”-first perspective: The reader is kept at the center.
• Conversational tone:  Never a hard sell or a desperate meeting request.

Call Question Generation

A third GPT-based prompt steps in, generating tailored discovery questions to guide the initial sales call, and helping reps dive into relevant pain points based on the lead’s business model and our value proposition.

All the rep needs to provide is the company name, website, and contact name, and the system takes it from there, turning a few data points into a structured, insight-driven conversation.

With just a few clicks, SDRs go from blank page to fully prepared: scored lead, crafted email, and tailored call questions in hand.

Results

While this tool is still in internal testing and has not yet produced external metrics, early feedback from our sales team is highly positive:

• Time savings: 60+ minutes preparation per lead now takes just a few minutes, including scoring, email creation, and call prep.
• Improved lead selection: Reps report more confidence in choosing which leads to pursue based on objective, repeatable scoring criteria.
• Higher engagement: The emails are described as “well-written,” “more personalized,” and “more likely to get a response,” though formal A/B testing is still pending.
  
  

About the Client

The tool was developed by SmartCat’s Labs team to support our own SDRs, who previously relied on manual research and outreach processes. 

This internal rollout not only allowed the SmartCat Labs team to experiment with cutting-edge technologies, but also test and refine the tool in a live environment with real leads, before considering external release.

Technologies Used

To create the AI scoring prospector, we relied on the following technologies:

• GPT-4.1 with web preview: For lead scoring, personalized emails, and question generation.
• Python: Backend logic and orchestration
• Web application: Interface for input and output
• No external APIs or third-party enrichments: These werent; used at this stage.

Your teams can work smarter. See how SmartCat’s AI agent solutions turn can everyday workflows into autonomous, efficient systems or book a strategy call with us!