Logistics Optimization

Logistics transformation at a renewable energy provider

Energy & Utilities

A custom-built, cloud-based Deep Learning solution that extracts data from paper-based receipts, was used to automate the client’s manual Order to Cash process. A phased approach set up the client for 3X growth without the corresponding need to add resources.

Client

Leading supplier of timber & biomass products in North America and Europe
Provides seamless transactions from inception to delivery
Offer door-to-door delivery via truck, railroad and barge

Business Challenge

Client provides door-to-door delivery of timber and biomass products
Manage logistics from inception to delivery, including sourcing, procurement, carrier logistics, tracking, billing, etc
Was finding it difficult to keep up with the complexities involved in handling nearly 1000 truck shipments per month, primarily via phone, email and manual billing
Expected to grow to approx. 3000 truck shipments per month by 2022
Had to manage challenges in finding and training back office resources to keep up with increasing workload

Our Approach

We commenced with discussing client’s mid-term expansion goals with leadership team
We then mapped out current as-is process and identified gaps, in order to address immediate workload needs
In the first phase, we setup Deep-learning based OCR to recognize and extract structured data from paper-based receipts to eliminate manual data entry
We connected Contracts, Orders, Shipments, Scale Tickets and Invoicing together in an integrated cloud-based custom-built platform that greatly reduced emails and phone calls to source raw material, obtain supplier forecasts, schedule trucks, send invoices, etc
Additionally, we also laid out a phased approach to distribute workload away from the client to some of the suppliers, carriers and truck drivers (while identifying means to incentivize these additional parties)

Results

Administrative workload reduction from ~25 hours per week to less than 15 mins per week
Created ability to capture meta-data on transactions that will allow additional big-data analytics

Key Takeaways

Our team custom-built a deep learning based cloud platform to perform data extraction from paper-based receipts. This allowed us to automate a large part of the client’s Orders to Cash process. We developed a roadmap to allow the client to scale 3X without hiring additional back office resources. The extracted data has also provided the ability to optimize sourcing, pricing, procurement, distribution and routing.