Enterprise AI Solution

GOAT - Gartner's Own Agentic Tech

Conversational Agentic AI Assistant for Gartner Associates

March 2024 - Present • 1 year 9 months

Company

Bitwise Solution Pvt Ltd

Client

Gartner

Status

Ongoing

Architecture

Serverless AWS

Executive Summary

GOAT (Gartner's Own Agentic Tech) is a conversational AI assistant designed to support Gartner associates by providing quick access to client information, engagement data, and research materials. Built on serverless AWS architecture, it streamlines workflows and enhances productivity through intelligent multi-agent orchestration and natural language interactions.

Business Justification

The Conversational Agentic AI Assistant is a digital tool designed to support Gartner associates. Its primary function is to streamline and enhance the efficiency of associates by providing quick access to relevant information and resources. This tool aims to improve productivity and ensure that associates are well-prepared for client interactions by offering comprehensive data retrieval and search capabilities.

What It Does

Client Information Retrieval

  • Fetch detailed client information including Major Client Priorities (MCPs) and Critical Issues (CIs)
  • Provide insights into past engagements and interactions with clients
  • Help associates stay informed and prepared for client meetings

Call and Engagement Data

  • Retrieve transcripts and follow-up emails from previous calls
  • Ensure continuity and context in ongoing client relationships
  • Access historical engagement data for comprehensive client journey view

Document and Research Search

  • Search Gartner's extensive repository of documents and research materials
  • Provide associates with the latest insights and data
  • Support decision-making with relevant, up-to-date information

Workflow

The GOAT system follows a streamlined workflow designed for ease of use and efficiency:

  1. 1

    Okta Login

    Associates log in using Okta authentication for secure access

  2. 2

    Agent Selection

    Select the client/account to work with

  3. 3

    Setting Up Tool

    Configure the tool with the selected client details

  4. 4

    Ask Query

    Associates can ask queries or prompts in natural language

  5. 5

    Get Response

    The tool provides intelligent responses based on the query

Data Model

The GOAT system uses a comprehensive data model to manage user access, agent configurations, and interaction history:

Emp-Agent-Relation

Controls user-agent access and permissions

Agent-Network-Junction

Manages primary and secondary agents for multi-agent orchestration

Agent-Repository

Stores agent configurations and context information

Agent-Tool-Junction

Links agents with available tools and capabilities

Tool-Box

Contains tool configurations and query templates

History

Tracks session and engagement data for continuity

Feedback

Captures user feedback for continuous improvements

Architecture

GOAT is built on a serverless architecture leveraging AWS services for scalability, reliability, and cost-efficiency:

Why Serverless?

Reduced Operational Overhead

No server management—AWS handles all infrastructure, scaling, and maintenance for Lambda. Previously, EKS caused timeouts and blocking issues with many concurrent users.

Automatic Scaling

Lambda automatically scales to handle any number of requests without manual configuration. Eliminates bottlenecks by instantly scaling to meet demand.

Cost Efficiency

Pay only for actual usage with Lambda (invocations and compute time), unlike EKS where cluster resources incur costs even when idle.

Enhanced Security

Lambda offers fine-grained IAM permissions and isolated execution environments, enhancing security and compliance.

Performance Comparison: EKS vs Serverless

Testing with 200 concurrent users over 60 minutes showed dramatic improvements:

29,746

API calls processed with Lambda (vs 9,742 with EKS)

3x

More requests handled with serverless architecture

Technical Implementation

Technology Stack

PythonAWS LambdaAPI GatewayServerlessLLMAIAgentic AIPostgreSQLS3OktaREST APIMulti-Agent SystemConversational AIRAGVector Database

1. Multi-Agent Orchestration

Implemented intelligent agent routing and coordination:

  • Primary and secondary agent management for complex queries
  • Context-aware agent selection based on query type
  • Agent network junction for seamless handoffs
  • Tool-agent mapping for specialized capabilities

2. Conversational AI

Built natural language interface with LLM integration:

  • RAG (Retrieval Augmented Generation) for accurate responses
  • Vector database for efficient document retrieval
  • Context management across conversation threads
  • Session history tracking for continuity

3. Security & Authentication

Implemented enterprise-grade security measures:

  • Okta integration for SSO authentication
  • Role-based access control (RBAC)
  • Token-based API authentication with 2-hour expiry
  • Data encryption at rest and in transit

Security & Compliance

GOAT follows enterprise security standards and compliance requirements:

Identity and Authentication

OAuth 2.0 token-based authentication with Okta SSO integration

Data Encryption

Encryption at rest (AWS RDS, S3) and in transit (HTTPS/TLS)

Access Control

Network segmentation with EKS and fine-grained IAM permissions

Logging and Auditing

Comprehensive logging to Splunk for audit trails and monitoring

Backup Strategy

Automated backups for RDS (99.999999999% durability) and S3

Results & Impact

3x

Increase in API throughput with serverless architecture

60%

Reduction in response time for client information retrieval

200+

Concurrent users supported without performance degradation

Business Benefits

  • Enhanced Productivity: Associates spend less time searching for information and more time engaging with clients
  • Improved Client Interactions: Quick access to client history and insights enables more informed and personalized conversations
  • Scalable Architecture: Serverless design handles growing user base without infrastructure concerns
  • Cost Optimization: Pay-per-use model reduces infrastructure costs compared to always-on EKS clusters
  • Continuous Improvement: Feedback system enables iterative enhancements based on user experience

Transferable Learnings

The GOAT project demonstrates several key principles applicable to other enterprise AI initiatives:

Agents for Value Story Journey

Multi- agent systems can be adapted for various business journeys beyond client management

Prototype to Scale

Serverless architecture enables rapid prototyping and seamless scaling to production

AI-Driven Insights

LLM integration provides intelligent, context-aware responses that improve over time

Conclusion

GOAT (Gartner's Own Agentic Tech) represents a significant advancement in enterprise AI assistants, combining conversational AI with multi-agent orchestration to deliver intelligent, context-aware support for Gartner associates. The migration from EKS to serverless architecture resulted in 3x improvement in throughput while reducing operational complexity and costs.

The platform's modular design, comprehensive API, and robust security measures make it a scalable solution for enterprise knowledge management and client engagement. The feedback system and continuous improvement cycle ensure the assistant evolves with user needs.

Need Enterprise AI Solutions?

I can help your organization implement conversational AI assistants, build multi-agent systems, and architect scalable serverless solutions on AWS.

Let's ConnectView More Projects