AI Smart Assistant on AWS for effective customer support

Farby.sk, a leading Slovak retailer of paints, coatings, and accessories for DIY enthusiasts, transformed its customer support by deploying an AI Smart Assistant built on the AWS platform. The solution, developed in collaboration with Aspecta, has automated more than 60% of common requests, accelerated responses by more than 90%, and improved customer satisfaction by 12 CSAT points.

7. October 2025 ┃ 6 minutes reading

Challenge

With the rapid growth of its e-commerce channel, Farby.sk faced a surge in customer inquiries and support requests. This mirrored a wider retail trend where the growth of online sales and rising expectations for instant communication put increasing pressure on customer service teams. Retailers across the sector struggle to scale support, maintain consistent service quality, and manage rising operational costs.

More than 65% of all incoming tickets were repetitive, low-complexity questions such as “Where is my order?”, “What’s the delivery time?”, or “How can I return a product?”. Handling these manually slowed down response times, increased costs, and limited the team’s capacity.

Key Challenges

Long response times (8–10 minutes) and unavailability outside business hours.
Low customer satisfaction (CSAT below 75%).
High operational costs due to seasonal staffing needs.
Lack of scalability amid growing online sales volume.

Solution

Aspecta designed and implemented an AI Smart Assistant powered by AWS Generative AI services, automating repetitive interactions and delivering instant, accurate, and consistent responses across all customer channels. This approach was chosen for its ability to combine high automation levels with minimal infrastructure management. AWS provided the right balance between scalability, security, and speed of deployment. Amazon Bedrock enabled the use of advanced language models without the need for custom training, reducing time-to-market and costs. The serverless architecture using Lambda and API Gateway ensured flexibility and easy adaptation for future system expansion.

Solution Architecture

Amazon Bedrock – foundation models (Claude/Titan) for multilingual, context-aware responses
Amazon OpenSearch (Vector Engine) – semantic retrieval of FAQs, policies, and order data
API Gateway + AWS Lambda – serverless orchestration between Bedrock, OpenSearch, and ERP/e-commerce systems
Amazon DynamoDB – session context and conversation management
Amazon S3 – centralized repository for documentation, FAQs, and chat logs
CloudWatch, GuardDuty, CloudTrail – monitoring, security alerts, and auditing

Integration with Farby.sk Systems

The existing ERP and e-commerce platforms remained on the company’s infrastructure. Integration was achieved through a secure middleware layer providing authenticated, encrypted access to order, payment, and delivery data.

Implementation

The project followed an iterative three-phase delivery:

Analysis and Design – identification of the most frequent inquiries, domain prompt design, and API mapping
AWS Deployment – configuration of Bedrock, OpenSearch, and Lambda orchestration in a multi-AZ setup with CI/CD pipeline
Pilot and Optimization – A/B testing of responses, accuracy improvements, and multilingual support (SK, CZ, HU)

Overcoming Challenges

ERP Integration: Limited access to historical data solved via proxy API layer
Knowledge Base Updates: Combined event-driven reindexing with manual approvals for compliance
User Adoption: Pilot campaign and embedded chat interface directly integrated into the e-shop frontend

Implementation proceeded smoothly, without major issues. Leveraging fully managed AWS services and a well-prepared integration layer, the solution was deployed to production within a few weeks. The project achieved a rapid time-to-value and enabled immediate KPI tracking.

Results and benefits

Deployment of the AI Smart Assistant delivered measurable improvements within the first months of operation:

60 % automation of repetitive requests
>90 % faster response times (from 8 minutes to real-time answers)
+12 CSAT points improvement in customer satisfaction
24/7 availability without increasing staff
Reduced operational costs for customer support and seasonal hiring

The solution enabled scalable online support operations without expanding headcount, creating a foundation for personalized digital assistance and sustainable growth.

Conclusion

The implementation of the AI Smart Assistant for Farby.sk demonstrates how Generative AI on AWS can effectively merge technology and customer experience. The project proved that intelligent automation can significantly reduce costs, improve service quality, and enhance satisfaction — all without disrupting existing systems.

This case also highlights the broader potential for automation in the retail and e-commerce sectors, where high interaction volumes present both challenges and opportunities. Similar AI-driven solutions empower companies to scale efficiently, ensure consistent customer communication, and refocus human resources on higher-value interactions. potenciál automatizácie v maloobchodnom a e-commerce segmente, kde vysoký objem zákazníckych interakcií predstavuje výzvu aj príležitosť. Nasadenie podobných riešení umožňuje firmám lepšie škálovať služby, poskytovať konzistentnú kvalitu odpovedí a sústrediť ľudské zdroje na hodnotnejšie úlohy.

Farby.sk’s success serves as a model example of how AWS Generative AI can drive efficiency, availability, and customer-centric innovation across the retail landscape.

Subscribe for newsletter