AWS re:Invent 2025 - Hesitate and You’re History: Why You Need Agentic AI Right Now (AIM246)

Summary of AWS re:Invent 2025 - Hesitate and You're History: Why You Need Agentic AI Right Now (AIM246)

The Importance of Agentic AI in a Data-Driven World

• Businesses today are driven by a multitude of decisions, both tactical and strategic
• Making informed decisions requires leveraging all available data, but many organizations struggle with data quality, accessibility, and actionability
• Failing to effectively utilize data can lead to costly mistakes and missed opportunities, as seen in examples of AI failures and social media blunders

Click's Approach to Enabling Agentic AI

• Click's core capabilities focus on three key areas:
  1. Providing trusted data products by extracting, integrating, and preparing data from diverse sources
  2. Uncovering hidden patterns and insights in data through advanced analytics
  3. Enabling organizations to act on data-driven insights with confidence

Ingersoll Rand's Journey with Agentic AI

• Ingersoll Rand, a large manufacturing conglomerate, faced the challenge of leveraging AI and analytics across its complex, multi-brand ecosystem
• The company adopted a two-pronged approach:
  1. Workplace/personal AI: Driving AI adoption and literacy among knowledge workers
  2. Enterprise custom AI: Transforming key business functions like sales, field service, and customer experience

Bridging the Gap with Data Platform AI

• Ingersoll Rand identified a gap between the growing demand for complex, real-time data-driven decisions and their ability to deliver reliable, secure, and scalable data infrastructure
• To address this, they partnered with Click to implement a "Data Platform AI" approach, leveraging their existing investments in data and analytics

Key Components of Ingersoll Rand's Agentic AI Architecture

1. Data Sources: Integrating data from 50+ ERP, CRM, and other legacy systems
2. Extraction Layers: Enabling near real-time data movement and transformation
3. Medallion Architecture: Maintaining a curated "Goldilocks" layer of high-quality data
4. AI Factory: Scaling the deployment of reusable, reliable AI agents (MCPs)
5. Agentic Interfaces: Empowering users and other AI agents to interact with the data platform

Benefits and Outcomes

• 20,000 hours saved for account managers by automating data-gathering tasks
• Order management teams in Asia-Pacific able to access a single dashboard, eliminating the need to navigate multiple CRM and ERP systems
• Combining descriptive, diagnostic, predictive, prescriptive, and perceptive analytics in a unified platform

Recommendations and Takeaways

1. Leverage your existing data and analytics investments as a foundation for agentic AI
2. Empower your analytics "super users" to become an instant AI army
3. Prioritize data governance and feedback loops to ensure trustworthy AI-driven decisions
4. Embed analytics and AI capabilities directly into business workflows and processes
5. Adopt an enterprise-grade, adaptable approach to data infrastructure and agentic AI

Technical Details and Metrics

• Iceberg-based data platform supporting 17 million transactions per second
• Leveraging technologies like Anthropic, Microsoft Copilot, and Click's own agentic framework

Business Impact

• Increased productivity and efficiency in key business functions like sales, field service, and customer experience
• Improved decision-making by providing real-time, contextual insights across the organization
• Reduced costs and risks associated with data quality and accessibility issues

Real-World Examples

• Automating 20,000 hours of manual data-gathering tasks for account managers
• Enabling Asia-Pacific order management teams to access a single, unified dashboard instead of navigating multiple legacy systems