💪🏻 This was the project that I (Akash Bulbule) and my fellow PM ( Monika) were responsible during our Journey in ChargePoint, Aimed to increase station owners revenue

EV Industry context

About ChargePoint

Why this project?

1. It solves a High-Cognitive Load problem (Complexity → Clarity)

2. Demonstrates AI as a Utility, not just a prototype

3. Connects Design to Business Levers (EBITDA)- Design moving Business needle

Project Content:

• Context on EV industry (ChargePoint)

• Problem / Data insights / Pod structure

• Defining scope for project

• Competitors & Research

• Approach & Collaboration

• Solution screens

• Conclusion

Team Structure

• Product lead: Monika (Sharfuz)

• Data & AI lead: Ravi, Federica

• Engineering lead: Sunny

• UX lead: Akash B

Problem Statement🕵🏻‍♂️ 

• Majority of EV charging stations are underperforming financially and are not profitable for station owners, which creates problem for future infrastructure investment from

• Why?
  Coz Many major EV charging networks currently Use manual static pricing or a one-size-fits-all approach based on average state energy costs and basic competitive prices.

  
  Problems in detail

  • Set pricing are only banked on few inputs (Demand trends, Cost variability, and Optimal revenue opportunities)

  • Most public Sites sit under-utilised at slow hours, get slammed at peak hours, and pay punishing demand charges

  • Station Hosts are not as savvy or do not want to constantly monitor & updating charging prices based on competitors

  • Hosts fear what if drivers don’t show up if price increased

Data Insights

Type of station owners

Competitors study

Our goal

Enable Electron sellers to maximise revenue by AI-driven pricing recommendation engine to improve pricing changes

How are we doing it?

Introducing Smart pricing AI engine in our Existing Pricing Flows

What is Smart pricing AI engine

AI model helping CPO’s move beyond static pricing for our EV charging stations.

• In stead of "Set it and forget" price, this engine provides Dynamic recommendations that adapt to real-world conditions, aiming to Maximise revenue while remaining competitive and fair to our customers.

♵ My role as Product Designer

• Collaboration: Worked at the intersection of Product, Engineering, and Data Science to define core workflows.

• User Research: Facilitated PM interviews to identify revenue-generation pain points for owners.

• AI-Driven Ideation: Developed rapid concepts for quick validation of the AI pricing recommendation engine.

• Validation: Created interactive prototypes for stakeholder syncs and conducted rigorous user testing to ensure design efficacy.

Let's relook on problems

Problems

EV charging stations are not profitable coz they use manual static pricing or a one-size-fits-all approach based on average state energy costs and basic competitive prices.

Problems in detail:

• Set pricing are only banked on few inputs (Demand trends, Cost variability, and Optimal revenue opportunities)

• Most public Sites sit under-utilised at slow hours, get slammed at peak hours, and pay punishing demand charges

• Station Hosts are not as savvy or do not want to constantly monitor & updating charging prices based on competitors

• Hosts fear what if drivers don’t show up if price increased

How solve it?

1. Build Data-driven recommendation engine for pricing.

2. Leverage AI/ML where feasible to adapt to real-world conditions

3. Optimise for:

   • Revenue maximisation for station hosts (CPOs).

   • While remaining Competitive in the market (Eliminating Monopoly)

   • Fairness for loyal customers/drivers (Trust)
     - With Station Hosts: Dynamic pricing must be transparent, explainable, and customisable. Hosts should never feel coerced or out of control.
     - With Drivers: Drivers should understand when and why prices change. We will use consistent pricing explanations to ensure perceived fairness.
     - With Our Ecosystem: We will never expose or use one host’s pricing to influence another host's pricing. Data privacy and competitive integrity are paramount.

Defined Metric Goals

• Core metric: Site Revenue (Target = >5% lift) (Pure profit)

• Counter metric: Site Utilisation (Target = >25%) (How many hours in a day is my station in use)

Factors Driving Pricing

User story based on Interviews conducted

Solution 🌊

Hybrid Engine

Smart pricing would be a hybrid engine which combines rule-based components with machine learning (ML) models to forecast utilisation based on historical and contextual data.

• Rule engine recommends optimal pricing scenarios.

• ML models then forecast expected revenue from these recommended prices.

External (Electricity cost)

• Option 1-Direct text Input of utility rate manually through a form

• Option 2- Pre-filled Estimation: Pull regional rates via OpenEI API

• Option 3 - Upload a PDF/CSV of the bill to identity Utility provider & Plans

  

Internal Data inputs (Site level)

• Utilisation (sessions/time of day, occupancy)

• Energy dispensed (kWh/session, peak kW)

• Pricing history & previous experiments

Internal Data inputs (Site level)

• Historical transaction data (10+ years available) - CP & Roaming

Key Inputs to the Engine by Data team

• Location Quality (Location Score):

  • A critical factor influencing driver choice alongside price and charging speed.

  • Stations near attractive amenities (restaurants, cafes, shops) have higher value.

• Location score calculation:

  • Uses Google Places API data of businesses around the station.

  • Classifies businesses by category and incorporates their Google ratings.

  • Accounts for proximity (distance) to the station from amenities

  • A generative weighting model combines these factors to produce a numeric location score.

The location score serves as an important input feature in the pricing and utilization forecasting models to better tailor recommendations based on site attractiveness and competitive landscape.

Key Ingredients: What Data We Use

The engine makes its decisions by analysing data points providing a unique insight into a station's performance and market position.

Data Team Prototype:

Figma Make Prototype:

Problems v/s Solution

Outcome ⚡️⚡️

Successfully validated the impact of the AI Pricing Engine through direct engagement with 3 key customers with 7% revenue increase in 3 weeks timeframe

Summary & Strategic Pivot

Collaboration

My collaboration with the Product Management team continued till March 2026, during which I spearheaded the early-stage ideation, core feature definition, and the foundational workflow architecture.

Organisational Restructuring,

In late March 2026, the organisation underwent a major restructuring that led to a shift in product priorities.

Validation

Despite the project being placed on hold, we successfully validated the impact of the AI Pricing Engine through direct engagement with 3 key customers via detailed calls and data analysis.

Technical Challenges

• The pivot was further influenced by dependencies on a new data architecture.

• The AI recommendation engine required precise "site definitions" to function. However, the manual migration of users from the legacy system to the new architecture took longer than anticipated. This underscored the importance of data readiness in AI-driven design.

Future plan

The qualitative feedback from 3 customers confirmed that our design direction was sound, even as the implementation was rescheduled for future quarters.

Conclusion

• While the company later shifted its strategic focus, this case study serves as a definitive record of the vision, solutions, and design systems I established during this phase.

• It captures a robust design framework built to solve complex pricing and editing challenges in a competitive market.