Case Study(Academic Project)

AI-Driven Returns Optimization — Scheduling & Cost Case Study

Improving Amazon’s returns workflow using AI forecasting, structured scheduling, and cost-efficient planning.

PROJECT OVERVIEW

This project focused on designing an AI-powered returns optimization system to streamline Amazon’s high-volume product return operations. Our objective was to reduce processing time, minimize inventory congestion, and enhance customer satisfaction using predictive analytics.

My role was centered on:

developing the project schedule

defining activity breakdowns

estimating costs

establishing baselines

performing earned value analysis (EVA)

ensuring realistic and achievable timelines

This case study demonstrates how disciplined scheduling, forecasting, and cost control can guide a complex technical project toward successful delivery.

Details

Delivering a High-Complexity AI Project with a Clear Schedule & Cost Control

Challenge

Amazon’s returns process faces critical operational issues:

  • High return volume leading to delays and congestion

  • Inefficient manual sorting and routing

  • Lack of forecasting for peak return periods

  • Rising operational costs due to reprocessing delays

  • Limited visibility into real-time return patterns

The challenge was to design a scalable solution and create a realistic, structured PM schedule and cost framework for development and deployment.

Solution

We applied PMBOK-aligned scheduling and cost management techniques to define a clear roadmap:

  • Work Breakdown Structure (WBS) to clarify project scope and deliverables

  • Activity definition + sequencing using dependencies and logical flows

  • Duration estimation using expert judgment and parametric estimation

  • Cost estimation across hardware, software, labor, and integration

  • Schedule creation using Gantt charts + CPM

  • Baseline development to lock scope, timeline, and costs

  • Earned Value Analysis (EVA) to track schedule and financial performance

This structure transformed complexity into a manageable plan with predictable outcomes.

WORK PACKAGES

12

PROJECT DURATION

182 

COST BASELINE

$44k

CRITICAL PATH ACTIVITIES

7

A Structured Framework for Scheduling & Cost Control

We used a disciplined, PMBOK-aligned approach to convert a complex AI initiative into a clear execution plan. The focus was on task clarity, dependency management, accurate estimation, baseline creation, and performance monitoring through Earned Value Analysis (EVA).

1. Identifying Major Work Components

We began by outlining the major activities required to deliver the AI-powered returns system, including AI model preparation, data integration, backend alignment, user interface development, testing, and deployment support.

2. Establishing the Schedule & Cost Baseline

We developed the initial time and cost expectations for all major project activities. This included defining milestones, estimating duration, and allocating budget across development, testing, and deployment.

— Shows timeline from April → October
— Shows cost commitments
— Perfect for this section

3. Earned Value Analysis (EVA) for Performance Tracking

We applied EVA to evaluate the project’s cost and schedule health using PV, EV, and AC. Through CPI and SPI, we assessed how closely the work was aligning with the plan.

— This is your key performance visual
— Shows project early variance clearly

4. Analyzing Performance Variance

We examined the deviation from the baseline to diagnose schedule delays, resource bottlenecks, and cost overruns. This formed the basis for strategic corrective actions.

— SPI = 0.85 (behind schedule)
— CPI = 0.89 (over budget)
— Shows variance visually

5. Identifying Cost Overruns

We analyzed which phases contributed most to budget variance. Data processing delays, integration complexity, and extended testing cycles were the major cost drivers.

— Displays cost variances by component
— Use as a single crisp visual

6. Revising the Plan & Establishing a Recovery Strategy

Using performance insights, we created a revised timeline and new cost projections. This allowed us to guide stakeholders toward a feasible path forward.

— Shows new finish date
— Shows revised budget

Details

Results & Reflection

Outcomes

The project delivered a complete, data-driven scheduling and cost management framework for Amazon’s AI-powered returns system. Key outcomes included:

  • A validated schedule baseline covering all major phases of the project

  • Clear cost projections for development, testing, integration, and deployment

  • Earned Value Analysis (EVA) metrics to monitor cost and schedule performance

  • Identification of critical variances through CPI, SPI, and variance tables

  • A revised timeline and recovery plan, balancing time, cost, and resource constraints

  • Stakeholder-aligned corrective actions guiding the project back toward feasibility

These outputs transformed an ambiguous technical initiative into a structured, trackable execution plan.

​

Reflection

This project strengthened my capabilities as a Tech Project Manager, especially in roles requiring analytical decision-making and quantitative control. It helped me develop:

  • Precision in scheduling for complex, high-dependency technical projects

  • Confidence in cost estimation using structured PMBOK-aligned methods

  • Strong understanding of EVA to communicate performance objectively

  • Skill in baseline creation, allowing clear measurement of progress

  • Ability to diagnose variance through data rather than assumptions

  • A builder’s mindset, breaking technical uncertainty into measurable, controlled steps

Through this experience, I gained a deeper appreciation for how schedules, costs, and performance indicators shape real-world execution — especially in AI and automation projects. These foundations directly support my long-term goal of becoming a strong Tech Project Manager in digital and AI-driven environments.

​

WORK PACKAGES

12

PROJECT DURATION

182 

COST BASELINE

$44k

CRITICAL PATH ACTIVITIES

7

Quote

“Agility is the ability to adapt and respond to change… Agile is about thinking differently, being open to change, and finding ways to deliver value faster.”

Jim Highsmith (Co-Author of Agile Manifesto)

Get In Touch

Adivarekar.s@northeastern.edu

https://www.linkedin.com/in/swapnilaadi/