984: Building AI Agents Where 99.9% Accuracy Isn't Good Enough, with Raju Malhotra

In this episode of the Super Data Science Podcast, host Jon Krohn interviews Raju Malhotra, Chief Product and Technology Officer at Certinia, a leading provider of professional services automation software. Malhotra dismantles the 'SaaS-pocalypse' narrative, arguing that SaaS is not dying but evolving into a powerful hybrid model combining SaaS with agentic AI capabilities. He illustrates this transformation through real-world examples, such as Certinia’s AI staffing agent used by Siemens, which optimizes thousands of resource assignments in real time—tasks impossible for humans alone. The conversation dives deep into how Certinia leverages Salesforce’s AgentForce 360 platform to build enterprise-grade AI agents that operate with both probabilistic intelligence and deterministic guardrails, ensuring compliance and accuracy critical in financial contexts. Malhotra emphasizes that 99.9% accuracy isn't enough in high-stakes domains like accounting, where even small errors can have serious consequences. He also shares insights on data quality, continuous experimentation, and the importance of a growth mindset for AI practitioners. The episode concludes with a book recommendation: 'Project Hail Mary' by Andy Weir, praised for its scientific rigor and thrilling narrative.