All Case Studies

Case Study — AI / Energy Tech

AI-powered electricity bill analysis for the Polish energy market with proprietary algorithm

5 Energy Operators
20+ Tariff Groups
4 PV Scenarios
<3 min Bill to Report

The Client

Breeze Voltuga is a Polish energy consultancy with a proprietary mathematical model for analyzing electricity bills and calculating photovoltaic investment ROI. The model originally lived inside MATLAB scripts that required an engineer to run manually.

They needed to turn that proprietary algorithm into a scalable, consumer-facing SaaS platform where anyone can upload a single electricity bill and get a full financial analysis — complete with tariff optimization and PV investment simulations — within minutes.


The Problem

Poland's electricity market is served by five major operators with a labyrinth of tariff groups. Most consumers have no practical way to determine if their tariff is optimal or what PV installation would actually save them. The client had a proprietary algorithm that solved this but it lived in MATLAB scripts requiring manual execution. The goal: turn it into a scalable SaaS accessible to anyone.

Challenges & Solutions

Challenge

Research-Grade Algorithm

The proprietary MATLAB algorithm needed to be re-engineered into a production-grade backend service capable of handling concurrent analysis requests.

Solution

Java Spring Implementation

Re-engineered the MATLAB algorithm into a Java Spring backend service, modeling full 24-hour x 12-month consumption matrices from a single bill using RDN market data as a demand proxy.

Challenge

Complex Tariff Landscape

Five operators, 20+ tariff groups (residential and commercial), each with different pricing structures — flat, time-of-use, and dynamic.

Solution

Comprehensive Tariff Engine

Built a tariff engine covering all residential (G11, G12, G-Dynamic) and commercial tariffs with hour-by-hour cost calculations including all fee components.

Challenge

Multi-Format Bill Ingestion

Bills come as PDFs, phone photos (JPG, HEIC), and scans (PNG) — each requiring different data extraction approaches.

Solution

OpenAI-Powered OCR

Integrated OpenAI vision/OCR to automatically extract consumption data, billing period, operator, and tariff from uploaded documents in any format.

Challenge

PV Investment Complexity

Multiple photovoltaic scenarios with different economic outcomes — net-metering, net-billing, dynamic tariffs, and storage options.

Solution

Four-Scenario Simulator

Implemented four PV investment scenarios: standalone with net-metering/net-billing/dynamic, hybrid storage, virtual storage (Fake PV), and zero-export — each with payback period calculations.

Challenge

Technical Results for Non-Technical Users

Complex energy analysis results need to be understandable by households and small business owners without technical backgrounds.

Solution

AI-Generated Reports

OpenAI generates plain-language explanations translating tariff comparisons and PV payback calculations into actionable recommendations with downloadable branded PDF reports.

What We Delivered

01

AI Analysis Engine

Proprietary energy algorithm + OpenAI integration for bill OCR, consumption modeling, and AI-generated report delivery.

02

Tariff Comparison System

Hour-by-hour cost calculations across all 5 Polish operators and 20+ tariff groups.

03

PV Investment Simulator

Four-scenario photovoltaic modeling with payback periods, savings projections, and storage simulations.

04

SaaS Platform

Angular frontend with user dashboard, admin panel, Stripe pay-per-analysis billing, and PDF report generation.

05

Ongoing Maintenance

Infrastructure management on DigitalOcean, tariff table updates, OpenAI model upgrades, and priority support.

Technology Stack

Angular
Java Spring
PostgreSQL
Firebase
DigitalOcean
OpenAI API
Stripe
Cloudflare
Sentry
SendGrid

Results

5 Operators

All major Polish energy distributors covered with 20+ tariff groups analyzed per bill.

<3 min

From bill upload to full AI-generated analysis report with PV simulations.

4 Scenarios

Comprehensive photovoltaic investment modeling including storage and dynamic tariffs.

Frequently Asked Questions

What is Breeze Voltuga?

An AI-powered SaaS platform that analyzes electricity bills for Polish households and businesses, identifying optimal tariff plans and calculating photovoltaic investment ROI across all five major Polish operators.

Which Polish energy operators does the platform support?

All five major operators: PGE, TAURON, ENEA, ENERGA, and STOEN, covering residential tariffs (G11, G12, G-Dynamic) and commercial tariffs.

What technology stack powers Breeze Voltuga?

Angular frontend, Java Spring backend, PostgreSQL database, Firebase for auth, DigitalOcean hosting with Cloudflare CDN, OpenAI for AI analysis, and Stripe for payments.

How does the proprietary energy analysis algorithm work?

It models hourly consumption using RDN market data as a demand proxy, factors in seasonal and appliance coefficients, extrapolates a 24h x 12mo matrix from a single bill, then simulates four PV scenarios calculating payback periods and savings.

Who developed and maintains Breeze Voltuga?

Designed, built, and actively maintained by Marotino INC with ongoing support, infrastructure management, and continuous platform improvements.

Have a Complex Platform Idea?

We turn proprietary algorithms and domain expertise into scalable, AI-powered products. Let's talk about what we can build for you.