### Accelerating IT Support Resolution with Generative AI: The MPAC Journey
MPAC, Ontario’s trusted authority in property assessments, manages a staggering 5.7 million properties. Recently, the organization faced overwhelming challenges at their IT support desk, witnessing an influx of routine requests that drained resources and adversely affected productivity across the board. To address these pain points without compromising on security and governance—critical for public sector operations—MPAC embarked on an innovative journey to implement a round-the-clock IT support solution.
With the creation of the **MPAC IT Support Assistant**, built on their custom **MPAC Orchestrator** platform using **Amazon Bedrock** and **Amazon Web Services (AWS)**, the organization has transformed its approach to IT support.
#### Achievements of the IT Support Assistant
The deployment of the MPAC IT Support Assistant yielded remarkable results in its first three months:
– **Over 4,500 support requests** handled autonomously.
– An impressive **83.5% autonomous resolution rate** for routine IT inquiries.
– Average response times slashed to **3.6 seconds** for employee queries.
– **Over 96% reduction in operational costs** due to serverless architecture.
– **Around-the-clock availability**, ensuring no more wait times for common support issues.
These achievements underline how effectively MPAC has harnessed generative AI technology to improve productivity.
### Solution Overview
The development of the AI-powered IT Support Assistant leveraged a serverless approach, powered by AWS. The key layers of architecture include:
1. **User Interface Layer**: Featuring a drag-and-drop interface for easy management of AI workflows.
2. **Orchestration Layer**: Built on **Amazon API Gateway** and **AWS Lambda**, it abstracts model management, prompt templating, data retrieval, and security behind a unified API.
3. **AI/ML Layer**: Employing **Amazon Bedrock’s** foundation models with vector database integration.
4. **Data Layer**: Utilizing **Amazon Relational Database Service (RDS)** for metadata, **Amazon S3** for document storage, and **Amazon Elastic Kubernetes Service (EKS)** for vector databases.
This architecture provides a solid foundation for rapid IT support, blending cutting-edge technology with user-centric design.
### Solution Deep Dive: The MPAC Orchestrator
The **MPAC Orchestrator** acts as the backbone of the organization’s AI capabilities, streamlining development and maintaining enterprise-grade governance. It wraps complex features including model management, prompt templating, data retrieval, and security into a singular, coherent platform accessed through a single API.
**Core Components of the MPAC Orchestrator**:
| **Component** | **AWS Services** | **Function** |
|———————-|———————————————————-|—————————————————|
| **Orchestrator API** | Amazon API Gateway, AWS Lambda (TypeScript), Amazon RDS | Manages users, endpoints, workflows, and audit trails |
| **Streaming Service** | Amazon EKS | Provides private streaming from Amazon Bedrock |
| **Data Ingestion** | Amazon SQS, AWS Lambda (Python), Amazon S3 | Processes documents with semantic chunking |
| **Security and Governance** | Amazon Cognito, AWS Lambda authorizer, Amazon VPC | Enforces single sign-on and access controls |
This structured approach allows MPAC to efficiently meet the growing demands for IT support while ensuring robust security measures are integrated within the architecture.
### Key Implementation Details
The implementation of the **MPAC Orchestrator** leveraged several innovative coding solutions to boost efficacy:
1. **Agentic Orchestration**: A supervisor pattern designed for dynamic tool execution ensures efficient task management, allowing for multi-step reasoning within a unified workflow.
typescript
// Sample TypeScript code representing the orchestration framework
export class AgenticWorkflow {
// Workflow execution route
}
2. **Built-in Cost Tracking**: Automatic tracking of token usage allows MPAC to maintain visibility on spending without requiring extra instrumentation.
typescript
// Lifecycle hooks tracking token usage
class AgenticWorkflow {
// Cost metrics collection
}
3. **Clean Lambda Handlers**: The orchestrator simplifies Lambda’s capabilities, allowing developers to focus on AI logic rather than infrastructure complexities.
typescript
// Simple Lambda handler for agent invocation
export const handler = awslambda.streamifyResponse(/* Handler Logic */);
### The AI Workflow in Action
The IT Support Assistant operates through an intricate agentic workflow that significantly enhances user experience by facilitating various paths:
– **Intent Entry Point**: Users are guided based on whether they seek general help or wish to escalate an issue formally.
– **Ticket Creation Path**: The system swiftly confirms issues, generates tickets, and communicates solutions effectively to users.
– **Self-Help Path**: Using advanced semantic analysis, the workflow extracts information from error messages and guides users with relevant documentation.
– **Advanced Question Routing**: Complicated queries are routed to live agents who are fully briefed on the context, ensuring quick and informed responses.
### No-Code Interface for Broad Accessibility
The **MPAC Orchestrator’s** user-friendly, no-code interface democratizes the development of AI solutions across the organization. It enables employees to upload unstructured documents directly into the vector database, facilitating seamless retrieval and processing. Users can create workflows using drag-and-drop functionalities, making advanced AI capabilities accessible to non-technical personnel.
With real-time analytics dashboards tracking token usage, costs, and performance, MPAC has empowered decision-makers to optimize resource allocation and drive efficiency.
### Future Roadmap: Expanding Capabilities
MPAC’s vision extends beyond IT support automation; it aims to integrate generative AI into various operational sectors such as human resources, facilities management, and beyond. The organization’s roadmap concentrates on embedding AI seamlessly within core business workflows, providing unified experiences across departments.
### Getting Started with Your Own Implementation
Organizations interested in following MPAC’s lead can begin by scheduling architecture reviews with their AWS teams and exploring the capabilities of **Amazon Bedrock** through the AWS Management Console. By leveraging the insights derived from MPAC’s structured implementation, organizations can balance innovation, compliance, and cost while enabling a successful transition to AI-driven operations.
With careful planning and execution, MPAC demonstrates how public sector organizations can not only innovate using generative AI but also ensure that deployment aligns tightly with institutional values such as security and service excellence.
### Real-World Results
MPAC’s commitment to adopting generative AI has significantly streamlined their IT support services, improved response times, reduced operational costs, and above all, enhanced user satisfaction. Their journey serves as a blueprint for other organizations aspiring to thrive in today’s rapidly changing technological landscape.
Ready to transform your IT support with generative AI? Engage with your AWS solutions architect to begin your journey towards innovation and streamlined operational excellence.
