Overview

Product / Initiative: Automated Data Scrubbing
Role: Principal Product Manager
Location: Global (Firmwide Initiative)
Scale: 20,000+ engagements annually, 5,000+ partners, global systems having sensitive client and pricing data across multiple geographies
Outcome: Automated the data scrubbing and anonymization process using AI/ML, eliminating manual effort, ensuring compliance, and reducing processing time by 85%, while maintaining security and confidentiality of client information.

Problem / Opportunity

When partners at firm created engagements, they frequently used highly sensitive client information, including pricing details, strategic priorities, and contractual numbers.

Key challenges:

🧾 Manual data scrubbing: Partners or support teams had to manually remove or anonymize sensitive fields before engagement data could be shared internally for approvals, pricing reviews, or analytics.
⏳ Slow and error-prone: Manual processes delayed setup, created inconsistencies, and introduced risk of human error.
🔐 Compliance risk: Sensitive data exposure during sharing and approval workflows created security and confidentiality risks.
📉 No standardization: Different teams scrubbed data differently, resulting in incomplete or inconsistent anonymization.

Opportunity:

• Automate data scrubbing and anonymization using AI to standardize, accelerate, and secure the process.

• Enable partners to share sanitized engagement data instantly and safely.

• Reduce operational burden, errors, and compliance risk.

Goals & Success Metrics

Primary Goal: Secure and accelerate engagement creation by automating sensitive data scrubbing with AI.

North Star Metrics:

• Scrubbing time per engagement

• Compliance error rate

• Partner satisfaction

Supporting Metrics:

✅ 85% reduction in data scrubbing time
🧠 100% standardization of anonymization logic
🔐 0 data exposure incidents during sharing workflows
📈 Faster engagement approvals and processing

Strategy & Approach

Vision: Build an AI-powered data governance layer that automates scrubbing and anonymization during engagement creation—ensuring speed, accuracy, and compliance.

• Automation First: Use NLP and ML models to identify and scrub sensitive fields (e.g., client name, pricing, contract values) automatically.

• Standardization: Apply uniform scrubbing rules across all engagements to remove human variability.

• Security by Design: Integrate the scrubbing engine into engagement workflows before data sharing.

• User Experience: Make the scrubbing experience seamless for partners—happening in the background without extra steps.

• Governance & Compliance: Ensure adherence to data protection policies and reduce risk exposure.

Frameworks Used

• RICE Prioritization to assess impact and feasibility

• Dual-Track Agile for iterative delivery with partner feedback

• AI/NLP models for entity detection and anonymization

• Data Governance Framework for compliance and risk mitigation

• Telemetry & Observability to track usage and error rates

My Role & Contributions as Senior PM

• Defined product vision and end-to-end roadmap for automating sensitive data scrubbing.

• Partnered with legal, security, and compliance teams to map sensitive data categories and design governance rules.

• Worked with AI/ML engineering teams to train and deploy NLP models for automatic entity detection and redaction.

• Designed seamless integration with engagement creation workflows, ensuring zero disruption for partners.

• Defined and tracked success metrics, including time saved, error reduction, and compliance adherence.

• Led change management and partner enablement, ensuring smooth adoption across practices.

Solution & Execution

🤖 AI-Powered Scrubbing Engine

• Automatically identified sensitive data (client names, pricing, contract details, financial figures) using NLP models.

• Applied standardized redaction and anonymization rules instantly during engagement setup.

• Enabled partners to safely share sanitized data with other functions.

🧭 Embedded Governance

• Integrated scrubbing engine directly into engagement creation workflows.

• Automated compliance checks and logging for audit trails.

• Reduced manual intervention to zero for standard use cases.

📊 Real-Time Processing

• Reduced scrubbing time from hours to minutes.

• Enabled instant downstream workflows for approvals and analytics.

• Added telemetry to detect and handle edge cases with low false positive rates.

🔐 Security & Compliance

• Ensured no sensitive information left the partner’s control without being scrubbed.

• Fully aligned with internal data protection and client confidentiality policies.

Impact & Results

⚡ 85% reduction in scrubbing time
🧠 100% standardization of anonymization logic
🔐 0 compliance breaches related to engagement data sharing
📈 Faster engagement approvals and time to go-live
😊 Significant reduction in partner and support team workload

Retrospective & Learnings

✅ What worked:

• Automating a highly manual, repetitive, and risky step had an outsized impact on both speed and security.

• Embedding AI invisibly into partner workflows drove fast adoption.

• Clear data governance rules upfront simplified ML model training and validation.

🛠️ What could be improved:

• Expanding scrubbing coverage to more nuanced contractual clauses earlier could have reduced manual exceptions faster.

• More interactive feedback loops from partners could accelerate model improvement.

🧠 Key learning:

Automation + AI applied to compliance-critical workflows can simultaneously boost speed, security, and trust — turning a governance bottleneck into a strategic enabler.