Case Study
AI-powered support delivers £1.7M savings and 47% faster resolution
About Westcon-Comstor
Westcon Comstor is a $4B global technology distributor operating in over 75 countries. They connect vendors, resellers, and service providers across the technology ecosystem, handling thousands of support tickets and sales inquiries every month.
The challenge
With large volumes of support tickets flooding in, Westcon Comstor's support teams were hitting bottlenecks. Manual triage and troubleshooting meant slow response times, high operational costs, and frustrated customers. Support staff were spending too much time on repetitive research, and escalations to senior engineers were eating up resources that could be better spent elsewhere.
The company needed a smarter way to handle support tickets without hiring more staff or sacrificing quality.
The solution
We embedded AI assistance directly into Westcon Comstor's existing Jira workflows using an event-driven, serverless AWS architecture. The solution uses Amazon Bedrock with Claude 3.7 and Nova Lite to power intelligent reasoning and rapid responses, orchestrated through AWS Step Functions, SNS, and SQS.
Here's how it works:
AI reasoning – Amazon Bedrock with Claude 3.7 handles complex problem-solving, while Nova Lite provides quick responses for simpler queries.
Intelligent search – Bedrock Knowledge Base combined with OpenSearch uses vector and hybrid search to find relevant solutions from historical tickets and documentation.
Smart agents – Bedrock Agents with MCP tool servers running on ECS Fargate can take actions, fetch data, and execute workflows automatically.
Seamless orchestration – Step Functions, SNS, and SQS coordinate the entire workflow, ensuring tickets are processed efficiently without manual intervention.
Reliable storage – Amazon S3 stores ticket data and AI-generated responses for future reference and continuous learning.
Full visibility – CloudWatch provides logs and metrics so the team can monitor performance and optimize the system.
Native integration – The solution plugs straight into Jira using REST APIs and webhooks, so teams keep working in the tools they already know.
The AI assistant analyses incoming tickets, suggests solutions based on historical data, and can even resolve common issues automatically. When human intervention is needed, it provides engineers with relevant context and recommended actions, cutting research time dramatically.
The results
The impact was immediate and measurable. Average ticket resolution time dropped from 4.2 hours to 2.2 hours – a 47% improvement. Escalations from Tier-1 to Tier-2 support fell from 35% to 23%, freeing up senior engineers for more complex work.
The serverless architecture means Westcon Comstor only pays for what they use, with no expensive platform migrations or lengthy training programs.
Westcon Comstor now has an AI-powered support system that learns and improves over time, delivering better service to customers while significantly reducing costs.
