Akshita Thakur - Case Studies

AI & Personalization

40% Adoption Lift for AI Personalization Platform

Cigna Healthcare | Patient engagement platform (15M+ members)

The Problem

Adoption of the engagement platform was stagnant at ~30%. Members were ignoring generic outreach, and clinicians were overwhelmed with manual follow-ups.

My Role

Sole Product Manager leading a cross-functional team of engineers, designers, data scientists, legal, and compliance.

The Solution & Trade-offs

Launched Propensity-Scoring Model: Implemented a Customer AI equivalent ML model to predict member needs.
Agentic AI Reminder Agent: Built an agent that auto-personalized channel (email/SMS/app) and timing based on user behavior.
Ruthless Prioritization: Killed three "nice-to-have" admin features to ensure we hit the 3-month launch target.

Impact
40% increase in personalized outreach adoption.
25% higher retention of high-risk members.
$18M annualized value (validated by Finance).

Key Learning: Speed of learning > speed of building. Weekly user interviews were the highest-leverage activity during this project.

Machine Learning & Operations

Predictive Gate Assignment ML Prototype

SITA | Airport ops platform for 400+ airlines

The Problem

Passenger stress was high during flight delays due to gates being assigned manually, leading to last-minute changes and poor wayfinding experiences.

My Role

Associate Product Manager — stepped up to own the initiative end-to-end.

The Solution

Rapid Research: Ran 40 airport staff interviews in just 2 weeks to understand the root cause of assignment friction.
Data-Driven Prioritization: Used RICE framework (scoring 380) to justify investment over other features.
Lean Execution: Partnered with one data scientist to ship an MVP in 6 weeks, stopping a competing "perfect" solution that would have taken 9 months.

Impact
22% reduction in passenger stress metrics.
Successfully rolled out to 12 international airports.

Key Learning: Influence without authority is built one relationship at a time.