Established in 2012, HES FinTech is now a provider of fintech solutions, such as HES LoanBox (Loan management system for banks, alternative lenders, and fintech companies), GiniMachine (AI-powered scoring and decision-making engine), and HES Collection Agent (smart software for recovery strategies).
12%
response rate from re-engaged leads
70%
Gini index for reliable lead scoring model
x5
reduction in manual workload
Challenge
Streamline sales workflows for better conversion
While HES FinTech’s marketing department was generating a significant volume of leads, the sales department was overwhelmed with manual work. Processing each lead required manual research, demos, and cost estimation, slowing response times and limiting scalability. Overall, sales efforts were spread too thin across low-value opportunities instead of being focused on high-potential leads.
After GiniMachine’s lead scoring, we spend less time qualifying leads in CRM, which allows us to allocate more time to pre-sales activities. Lead quality has significantly improved, and the average deal size has increased.
Approach
AI-powered lead segmentation and funnel optimization with GiniMachine?
To address the challenge, HES FinTech needed a way to segment leads for both the sales and marketing departments and then apply different strategies to these segments. Cold leads were supposed to be automated and approached with marketing materials, FAQs, and product demos, while highly rated leads were to be approached directly by the sales team.
Historical data from HubSpot (such as industry, employee count, the latest referral, contact data, etc.) that was gathered from submissions was sent to GiniMachine through an API. After data preparation, HES FinTech achieved a Gini index of 0.6, indicating that the model could be used for commercial purposes.
GiniMachine AI model predicted conversion likelihood based on behavioral attributes. The lead scores were returned by the model, and the outcomes were recorded in a designated field for lead distribution and reporting:
- Median quality score of all leads was 0.25, serving as the baseline for lead quality assessment.
- Leads with scores above the threshold were assigned to sales, while those below were filtered out.
- Low-quality and non-targeted leads received low scores (0.02), enabling their removal from the sales funnel or immediate automated processing using CRM tools.
- MQL and SQL leads were immediately directed to the sales department, improving the conversion rate from MQL to Customer.
Result
Efficient data-driven system for handling marketing leads
The scoring model has proved effective and showed immediate impact for operations and conversions:
- Sales focused only on high-value leads above the threshold
- Low-quality leads were filtered out or automatically nurtured
- MQL/SQL routing became instant and consistent
- Low-quality leads re-engaged through email campaigns showing a 12% response rate
Overall, lead handling shifted from manual efforts to automated, data-driven allocation with improved efficiency and conversion.
Download PDF version
