Context

Led delivery of an initiative, which integrated Palantir Foundry with Salesforce to create an AI-powered case management system. The platform leveraged machine learning for case clustering, priority labeling, and automated write-backs, enabling faster resolution cycles and improved data quality across multiple business units. The project involved building complex data pipelines, integrating enterprise platforms, and deploying AI-driven features to support daily operations.

My Role

Served as both Project Manager and AI Product Manager, with responsibility for execution oversight and product definition. As Project Manager, coordinated delivery milestones, managed cross-functional dependencies, and ensured compliance with enterprise deployment standards. As an AI Product Manager, I supervised functional requirements for case clustering and prioritization, translated business rules into automated workflows, and supervised how Palantir’s AI/ML capabilities were applied to operational processes. Managed the transition from development into production, facilitated knowledge transfer, and established a hybrid support model bridging BI teams and AMS.

Challenges

Integration complexity was a defining challenge, particularly connectivity issues requiring international escalation and multi-factor authentication obstacles. Memory constraints emerged during Salesforce’s ingestion of historical data, forcing pipeline restructuring. Team continuity was disrupted by high vendor turnover, while a non-standard deployment pathway bypassed typical AMS handover, creating ambiguity around long-term support. Scope creep also arose when stakeholders pushed for expanded functionality beyond the initial design, requiring careful management to preserve delivery timelines.

Approach

Structured delivery into phased milestones, focusing on foundational technical issues before production deployment. Implemented incremental data loads to mitigate memory pressure, designed scheduled write-back processes for real-time updates, and enforced comprehensive documentation standards for support readiness. Facilitated frequent cross-team meetings to align Palantir developers, BI, and AMS stakeholders. Positioned the application as a continuous improvement platform rather than a fixed-scope project, enabling iterative deployment of features and rapid user-driven enhancements.

Outcomes

Delivered a functioning AI-enabled case management system with core features: case clustering, priority labeling, and automated Salesforce write-backs. The platform processed 50–100 daily case summaries in production, with scaling capacity for historical data ingestion. Documentation and support materials were completed, enabling maintenance and iterative development. The hybrid support model allowed BI teams to continue agile enhancements while preparing for the eventual AMS transition.

Lessons Learned

• Flexibility in Process: AI/ML projects often require non-standard delivery approaches that balance agility with compliance.
• Early Issue Resolution: Addressing data connectivity and memory constraints early prevents downstream delays.
• Sustained Communication: Clear documentation and frequent alignment touchpoints are essential when managing complex integrations and high-turnover vendor teams.