Discover how SOAX achieved 80% customer support automation – while improving CSAT by 95% - through a layered generative AI system built for reliability and scale.

The expert: Kirill Markin, Former Head of R&D at SOAX.

The organization: SOAX, a mid-sized global data extraction platform helping businesses gather web data efficiently. The company focuses on scaling automation within data-intensive, user-facing environments.

The problem: SOAX needed to deliver 24/7 customer support while tackling scalability, repetitive queries, and resource strain -- without sacrificing the quality and speed of responses.

‍

Key decisions

Build layered reasoning and cross-referencing to reduce hallucinations and improve context
Embed AI workflows into daily operations
Select LLMs based on task requirements, not hype
Encourage tight, cross-functional collaboration between tech and support

Key results

80% automation of support queries
30–35% reduction in support hours
95% increase in customer satisfaction (CSAT)
Maintained human oversight for complex, high-context interactions

1. Interface versus Intelligence: Rethinking Chatbots in Real-World Applications

Why chatbots fail to deliver in real-world scenarios. Underscores the need to move beyond the chat-interface hype to smart and reliable solutions.

2. Who is SOAX?

The case study subject, SOAX.

3. The Pain-Point Drive to AI

The key pain-points that drove SOAX to adopt AI solutions.

4. The Solution: An AI-Powered, Smart Chatbot

The key functions and outcomes of SOAX's AI-powered, smart chatbot.

5. Development and Deployment

The pre-launch decisions, the tech stack, the iterative development approach, the build process, and the overall planning and strategy.

6. Ethical and Security Compliance

The measures implemented for ethical and security compliance.

7. Implementation and Integration (1.5-3 months)

How SOAX integrated its AI solution with existing workflows.

8. Operational Challenges and Solutions

The key challenges encountered by the team and how they addressed them, in problem-solution pairs.

9. Key Insights for Tech Leaders

Key findings of the case. Actionable insights for tech leaders looking to integrate a scalable and reliable customer chat-support system into their businesses.

10. Conclusion

Key insights with closing remarks.

‍

Meet the Experts

Kirill Markin

Former Head of R&D, SOAX

Kirill Markin brings over 12 years of experience in AI, data science, and business automation, with a proven ability to transform technical innovation into practical, impactful solutions. As a driving force behind SOAX's AI-powered customer support system, Kirill played a pivotal role in designing and implementing scalable architectures that balanced efficiency, adaptability, and cost-effectiveness. Kirill’s technical expertise includes prompt optimization, pipeline automation, and developing modular AI workflows to streamline operations. His innovative use of task-specific models and real-time feedback mechanisms ensured SOAX's AI systems remained secure, agile, and aligned with business goals. Describing himself as a “data monkey”, Kirill thrives on experimenting with emerging technologies and continually pushing the boundaries of AI-driven solutions. His hands-on leadership has solidified SOAX’s position as a leader in intelligent customer support automation.

Other case studies

Cart

$ 1.00

Generative AI for 80% Customer Support Automation (test)