Internal PM Case StudyEnergy Ops & LogisticsChevron (Hypothetical)

ChevronLift — Automated Crude & Product Scheduling Platform

ChevronLift is an internal scheduling cockpit for crude oil and refined products. It pulls together pipeline, terminal, and marine data to generate draft movement plans, highlight conflicts, and give schedulers a single, live view of tomorrow's barrel flows.

Prefer to see the product?

This page walks through the product thinking behind ChevronLift. If you’d rather click around the scheduling cockpit itself, you can jump straight into the live demo.

Open the ChevronLift cockpit

Primary Problem

Manual, fragmented scheduling

Schedulers stitch plans across emails, spreadsheets, and siloed logistics systems.

Who I Designed For

Schedulers & Traders

Crude/product schedulers, marine logistics, terminal ops, and desk traders.

My Role

Product Manager

Framed the problem, defined flows, and scoped a realistic scheduling cockpit with AI assist.

This is a portfolio artifact. Data and counterparties are illustrative, but the flows and constraints reflect how I'd approach a real scheduling platform for a major like Chevron.

Context

Scheduling is the nervous system of the barrel — and it’s held together with email

In trading and supply, "the barrels" only matter if they show up in the right tank, on the right vessel, at the right time. Today, crude and product schedulers are forced to reconcile pipeline nominations, terminal capacities, refinery run plans, and marine ETAs in tools that were never designed for the job.

Pain points today

  • Pipelines, terminals, marine, and refinery plans sit in separate systems — schedulers become the integration layer.
  • Daily schedules are built in spreadsheets that are brittle, single-user, and hard to audit later.
  • Conflicts (tank overfills, incompatible batches, missed laycans) are discovered late, not surfaced early.
  • Traders change their mind; schedulers carry the cognitive load of recomputing a complex puzzle on the fly.

What ChevronLift aims to do

  • Provide a single operational view of crude & product movements across pipelines, terminals, and marine.
  • Generate draft schedules based on current inventory, nominations, and constraints — schedulers stay in control, AI does the grunt work.
  • Detect and explain conflicts early so schedulers can act before traders and terminals feel the pain.
  • Make "what changed since yesterday?" and "what breaks if this vessel is late?" answerable questions.

Users

Who ChevronLift is built for

Scheduler

Owns the physical reality of the barrels.

Needs

  • A reliable picture of all movements planned for the next 1–7 days.
  • Early warning on conflicts before they become phone calls.

What ChevronLift gives them

A cockpit that auto-builds a draft plan, flags conflicts, and lets them override with context — not a black-box scheduler.

Trader / Trading Analyst

Owns the book, P&L, and optionality.

Needs

  • Confidence that physical constraints are surfaced before trades are locked in.
  • Fast answers to "what if we shift this volume or this laycan?"

What ChevronLift gives them

A shared view of the schedule and constraints so they can negotiate and restructure deals with eyes wide open.

Terminal & Marine Ops

Run tanks, docks, and vessel movements.

Needs

  • Clear view of what's coming and when.
  • Early warning on tank stress and dock conflicts.

What ChevronLift gives them

A filtered operational view tailored to their assets: tanks, docks, and movements at their sites.

Solution

A scheduling cockpit that keeps humans in charge and lets AI handle the recomputation

ChevronLift isn't a fully autonomous scheduler. It's a decision cockpit that keeps the humans in control of the book while offloading the heavy lifting of recomputing feasible plans when something changes.

Unified Movement Graph

Pipes, tanks, docks, and vessels represented in one mental model, not scattered across systems.

Draft Schedule Generator

Given current inventory, nominations, and constraints, the system proposes a feasible next-day plan that schedulers can tweak.

Conflict Detection & Explainability

Tank overfills, incompatible batches, dock conflicts, and late vessels flagged with human-readable justifications.

Scenario "What-if" Support

Schedulers can adjust volumes or timings and see how the plan recomputes in seconds instead of rebuilding a spreadsheet.

Flows

Key flows the live demo is built around

The prototype anchors on three flows that feel real to anyone who has lived in scheduling: tomorrow's plan, a late vessel, and a tank stress test.

1. Building tomorrow’s schedule

Goal: Generate a draft next-day schedule and sanity-check it.

1Scheduler selects region, system (e.g., Gulf Coast crude), and date.
2ChevronLift pulls in latest inventory, pipeline nominations, and planned receipts/shipments.
3AI generates a draft movement plan respecting tank minima/maxima and known constraints.
4Scheduler reviews conflicts and overrides specific moves as needed.

Outcome: The team starts from a consistent, feasible base plan instead of rebuilding it from scratch every day.

2. Handling a late vessel

Goal: Re-plan around a delayed marine arrival without breaking everything else.

1Marine ops updates a vessel ETA, triggering an alert in ChevronLift.
2System highlights movements and tanks impacted by the delay.
3Scheduler runs a "recompute" scenario that shifts certain pipeline or terminal moves.
4Resulting plan shows which actions reduce demurrage and tank stress the most.

Outcome: Schedulers can respond to late vessels with structured options instead of ad-hoc firefighting.

3. Tank stress test at a key terminal

Goal: Avoid overfills and minimum violations at a constrained site.

1Scheduler opens the terminal view for a high-traffic tank farm.
2ChevronLift projects tank levels over the next 7 days under the current schedule.
3System flags days where levels breach max/min thresholds.
4Scheduler adjusts timing or reroutes volumes; AI recomputes a conflict-free variant.

Outcome: Tank risk is caught days in advance, not when operators are already scrambling.

Demo

What the live prototype actually shows

Under the hood this is a simple Next.js app with mocked data, but the screens are designed to feel like something a Chevron scheduler could actually live in.

Movement Plan Board

Table-style view of planned movements (pipeline batches, tank transfers, loadings) with status, volume, timing, and conflict icons. Filters by system, date, and asset.

Conflict & Risk Pane

Side panel listing conflicts (tank overfill, incompatible batches, dock overlaps) with explanations and suggested fixes, powered by lightweight AI narration.

Scenario Sandbox

Small control panel where schedulers tweak volumes or timings and trigger a "recompute" to see an updated draft plan and tank forecast.

Outcomes

Target outcomes and how I’d measure success

Because this is a portfolio artifact, the numbers are directional. The measurement mindset is real: less chaos, fewer surprises, and lower P&L drag from operational friction.

Scheduling conflicts surfaced earlier

60–80%

Share of tank/dock conflicts flagged at least 24 hours before they would have been noticed in today’s workflow.

Demurrage exposure reduction

10–20%

Modeled reduction in demurrage driven by earlier response to late vessels and berth conflicts.

Time spent rebuilding schedules

-30–40%

Reduction in scheduler time spent reworking plans after a mid-day change.

Execution

How I approached ChevronLift as a PM (and where AI actually fits)

  • Started from real constraints — tanks, docks, pipeline batches, laycans — rather than "let's add AI to scheduling."
  • Defined human responsibilities first (scheduler owns the plan, trader owns the book), then used AI to recompute options and narrate risk.
  • Chose three anchor flows that feel painfully familiar to schedulers: building tomorrow, handling a late vessel, and managing a stressed terminal.
  • Scoped the UI to be believable as a Chevron internal tool — opinionated but not sci-fi, with synthetic data standing in for feeds from pipeline/terminal/marine systems.
  • Anchored on measurable outcomes like reduced conflicts and demurrage, not just "nice dashboards."

In a real engagement, this would sit alongside integrations to actual pipeline, terminal, and marine systems. For the portfolio, the goal is to make the product thinking and tradeoffs legible in a single, navigable experience.