The most underrated dataset in enterprise AI is your org chart | Rippling's Albert Strasheim

The podcast explores the transformative potential of AI in organizational efficiency, emphasizing the importance of understanding a company's structure - such as reporting lines and access permissions - as a foundation for effective decision-making and automation. A key theme is the use of Human Capital Management (HCM) and payroll systems as real-time, reliable sources of organizational data, which can serve as a "hook" to power broader business applications, including IT management, security, compliance, and AI-driven workflows. This trustworthy data enables dynamic automation, supports agentic engineering, and allows organizations to move beyond fragmented tech stacks by creating unified, human-centric platforms.

The discussion highlights how AI's impact can be measured and optimized in software development, with insights into engineering productivity, AI code reviews, and the evolution of the software development lifecycle. Building robust, scalable platforms requires foundational components like data layers, triggers, systems of record, and action systems, all integrated to support AI agents operating in close proximity to data. As AI introduces complexity and non-determinism, the podcast underscores the growing importance of evaluations (evals), observability, and subsystem monitoring to ensure reliability. Looking ahead, the vision is to develop intelligent data platforms - equipped with data catalogs, transformation tools, and AI-generated insights - that enable proactive, personalized business intelligence and position companies to compete in a data-driven future.

• What if you built a human-centric data layer for your SaaS app using real-time HCM data?

  • Move: Integrate with an HRIS like Rippling or BambooHR via API to pull live employee data (roles, departments, manager reports, status) and model it as a graph in your app. Use it to auto-provision access, assign workflows, or personalize features.
  • Why Now?: HCM systems are already maintained with high accuracy by HR teams - this data is reliable, always updated, and acts as a "Trojan horse" into enterprise workflows without requiring manual input from your users.
  • Expected Upside: Reduce onboarding friction by 70%+, increase stickiness through automated role-based experiences, and unlock enterprise sales by demonstrating IT/security compliance out of the box.

• What if you turned your app into an agentic workflow engine using org-aware AI agents?

  • Move: Build lightweight AI agents that use organizational context (e.g., who reports to whom, who has approval authority) to autonomously route tickets, escalate issues, or suggest decisions - starting with one high-friction workflow like expense approvals or incident response.
  • Why Now?: With reliable HCM-backed org data and improved LLM reasoning, agents can now act with contextual awareness - avoiding the blind spots that make generic AI assistants fail in real business processes.
  • Expected Upside: Cut operational latency by 50% in key workflows, reduce user effort, and differentiate your product as an "autonomous" solution versus passive dashboards or manual tools.

• What if you replaced fragmented internal tools with a unified employee graph powering multiple modules?

  • Move: Instead of building separate HR, IT, and finance features, create a core "employee graph" that connects all user actions, permissions, and attributes - then build lightweight apps (e.g., device provisioning, leave management, SSO sync) on top of it as modular functions.
  • Why Now?: Customers are fatigued by tool sprawl; platform thinking wins in the AI era. Starting with a unified data foundation lets you move faster, avoid silos, and respond to new automation demands with minimal rework.
  • Expected Upside: Achieve 3x faster feature velocity, enable cross-functional automation (e.g., offboard employee revoke access reclaim laptop notify finance), and increase LTV by bundling services naturally over time.

• Identify and integrate a core source of real-time organizational data (e.g., HR/payroll) as a foundational layer for automating access, workflows, and permissions in your software product.
• Build modular platform primitives (e.g., triggers, workflows, data objects) that enable AI agents to act securely and autonomously, starting with narrow use cases and expanding through experimentation.
• Focus on creating cross-functional "technology smoothies" by combining existing tools and data sources into unified workflows instead of building isolated features or waiting for perfect AI agents.
• Implement robust evaluation (eval) frameworks for AI components - test subsystems rigorously, not just end-to-end outputs - to ensure reliability in non-deterministic, agentic systems.
• Shift from narrow-feature development to platform thinking by designing for extensibility, allowing customers to build custom connectors, objects, and automations using your system as a central data and action hub.