A new way of working
To meet the demands of new customers, our member Keiro created an offering for embedding automation, scalability, and deep industry expertise into a high-performance platform solution. This case study explores how Keiro leveraged cloud innovation to streamline operations and enhance customer operations for their customers.
The challenge at hand
The core issue Keiro’s customers had been facing were scaling customer support and operational efficiency to meet growing demand across multiple countries, while maintaining agility for a diverse client base. Existing communication methods—like direct chats, phone calls, and emails—were fragmented and unsustainable, consuming valuable technical resources and leading to inconsistent customer experiences, slower response times, and missed opportunities. Without a scalable solution, the clients’ team risked being overburdened, limiting onboarding capacity and slowing business growth. Additionally, in a competitive market filled with emerging self-service advertising tools, there was a pressing need to evolve the platform with automation and intuitive campaign-building features to stay relevant and deliver a modern user experience.
The solution
To address these challenges, a bot-based application was developed using AWS-native services to ensure scalability, high availability, and fault tolerance. The architecture is built on the CrewAI framework, with agents powered by models hosted on Amazon Bedrock for semantic search and intelligent interaction. Data is stored in Amazon S3, with vector search capabilities enabled via Amazon Aurora Serverless and the pgvector extension. The application runs on Amazon ECS Fargate with autoscaling, and is deployed across multiple Availability Zones to ensure resilience. Cross-region deployment of Bedrock models ensures low-latency inference and redundancy. This setup enables continuous operation, dynamic scalability, and a robust infrastructure capable of supporting rapid growth and delivering a seamless user experience.
The path to success
Two key performance indicators (KPIs) have been identified to estimate the operational impact of this solution during its initial production phase:
-
Estimated Average Response Time to Customer Inquiries. Prior to deploying, customer questions were handled manually through a direct messaging format, with response times varying significantly. Due to the responder juggling multiple responsibilities, replies could take anywhere from several minutes to a few hours.
Outcome: The response is prone to be reduced since the response time will be immediate. The complexity of the questions will highly decide the reduction amplitude. -
Estimated Time Freed Up from Manual Inquiry Handling. The team member responsible for responding to inquiries previously spent several hours per week addressing repetitive or low-complexity questions without being able to address them all.
Outcome: As the customer transitions toward integrated automation, early indicators suggest a reduction in time spent on inquiries, particularly through highly personalized answers and improved user guidance. Full impact is expected once the support experience is enhanced with more AI agents or workflow-driven response layers in the whole campaign journey.
About Keiro
Keiro is a Scale member and is part of our AWS Migrations & IoT sub committees. Keiro helps businesses grow by making their digital operations smarter and more efficient. Whether it’s cloud solutions, automation, or data strategy, we work closely with our clients to turn technology into real business value. Our team combines technical expertise with a hands-on approach to deliver solutions that truly make a difference.