Modernizing Student Engagement at Zetech University with a Secure, Multi-Tier Agentic AI Chatbot on AWS
About the Customer
Zetech University is a leading, rapidly growing private university headquartered in Nairobi, Kenya. Operating in the higher education sector, Zetech University is committed to providing industry-aligned, technology-driven academic programs. Serving an expanding student base across physical campuses and virtual learning environments, the institution possesses the characteristics of a scaling SMB enterprise. The university has actively pursued a comprehensive digital transformation strategy, migrating its core applications to the cloud to enhance operational efficiency, automate service delivery, and elevate the overall student experience.
Customer Challenge
Zetech University experienced a massive influx of daily inquiries from prospective students, parents, and current students regarding admissions, course registrations, fee balances, and exam timetables. The university’s traditional manual support channels and rule-based systems were overwhelmed, causing response times to balloon past 24 to 48 hours during peak academic cycles. Information was siloed across various PDFs, student portals, and department-specific databases, leading to inconsistent answers and administrative friction. The short-term risk of not addressing this bottleneck was a degraded student experience, missed enrollment opportunities, and severe operational strain on IT and support staff. Long-term, the university faced rising operational costs from scaling support personnel linearly, reduced student retention, and reputational challenges in a highly competitive academic landscape.
Partner Solution
Coresys Solutions designed and deployed an enterprise-grade, stateful Agentic AI Chatbot powered by AWS. The solution features a modular, graph-based orchestration framework (inspired by LangGraph/Strands patterns) deployed on Amazon EC2 to manage complex, multi-step student workflows. The system leverages Amazon Bedrock (utilizing the Anthropic Claude model) as the primary reasoning and inference engine. To facilitate dynamic action execution rather than static FAQ replies, we built a standardized tool abstraction layer. This allows the Agent to securely interface with Zetech’s Student Information System (SIS) APIs and payment gateways to retrieve real-time registration and fee statuses.
For secure data retrieval, we built a RAG pipeline utilizing Amazon OpenSearch Service for semantic vector indexing and Amazon S3 for knowledge storage. Network traffic is protected at the edge by AWS WAF and routed to private subnets using an Elastic Load Balancer (ELB). Session state and multi-turn conversation context are managed with ultra-low latency using Amazon ElastiCache (Redis). To decouple workflows and protect backend databases during traffic spikes, Amazon SQS handles asynchronous message processing. To enforce security, we implemented role-based access control (RBAC) utilizing AWS IAM with the principle of least privilege, unified logging via Amazon CloudWatch, and centralized security threat detection with AWS Security Hub and Amazon GuardDuty.
95%
Improvement in response delivery Speed achieved.
75%
of student inquiries resolved entirely in self-service
$72k
Generated in AWS ARR with Amazon Bedrock representing 25% of consumption
Results and Benefits
The production launch of the Agentic AI chatbot transformed Zetech’s student services with outstanding quantitative outcomes:
- Immediate Latency Reduction: Average response times for complex, multi-turn inquiries were slashed from the 24–48 hour baseline to 1–2 seconds, achieving a 95% improvement in response delivery speed.
- High Automation and Resolution Rates: The agentic system successfully resolved 65% to 75% of student inquiries entirely in self-service, achieving an overall query resolution rate of over 85% without human agent intervention.
- Operational Workload Mitigation: The system drove a 40% reduction in manual support workload for Zetech’s administrative and ICT departments, freeing valuable personnel for specialized student support.
- Production-Grade Reliability: The solution generates an estimated $72,000 in AWS ARR (with Amazon Bedrock representing 25% of consumption), demonstrating a highly utilized, resilient, and enterprise-grade deployment that scales automatically during peak enrollment periods.
About the partner
Coresys Solutions Limited is a pioneering AWS consulting partner based in Nairobi, Kenya, specializing in cloud migration, application modernization, and advanced AI/ML solutions. We empower public sector organizations, educational institutions, and enterprises across East Africa to design, secure, and scale production-grade workloads on AWS. Coresys Solutions holds deep technical expertise in AWS-native security, serverless compute, and Generative AI architectures, serving as a trusted advisor for organizations seeking to achieve business-led digital transformation.