PALO IT Blog

Use Case: Transforming a transportation giant's digital product development with AI

Written by PALO IT | 24/09/25

As one of the most dynamic and technologically advanced sectors, transportation is experiencing steady growth, as evidenced by Forbes reporting 64% of businesses anticipating AI will significantly boost productivity. At a glance, the global AI transportation market size is expected to be worth around US$21.4 billion by 2033.

This growth is all well and good, but it necessitates change amongst businesses who strive to stay abreast of tech innovation and stand out from the crowd. In the highly competitive and customer-centric transportation industry, our client—a regional transportation leader—was facing long feature delivery cycles, varied code quality, and an increasing need to innovate at scale. The client partnered with PALO IT to redefine the way its digital teams built software. What follows is a step-by-step look at our approach.

Phase 1 – DevPulse: Measuring ROI of Generative AI Software

Context & Business Challenge

The business was grappling with slow adoption of its tools, having activated 170 GitHub Copilot instances but only achieving active engagement from 30-35 users. Additionally, it operated in a fragmented environment with numerous developer and project management tools. They aimed to track usage metrics across platforms like Bitbucket, JIRA, and SonarQube while remaining open to expanding their data sources in the future.

The company was particularly interested in monitoring metrics such as commit rate, lead time, change failure rate, and change cycle. However, they faced challenges with a high overall lead time due to numerous bugs in the developed code, which were flagged but often left unaddressed.

Our Solution

To tackle these challenges, PALO IT developed DevPulse, a DevOps analytics platform. The platform utilizes Apache Superset for dashboard visualization and is built on a flexible cloud architecture that supports Docker deployments and PostgreSQL databases. The engagement was structured as a Proof of Value (PoV) under the Azure Innovate Partner-led Dev Productivity program.

DevPulse integrated data from various DevOps sources, including:

  • JIRA, Bitbucket, CodeCommit, and SonarQube for development and quality metrics.
  • A NoSQL database for ingestion and storage.
  • Tableau dashboards for visualization and reporting.
  • GitHub Copilot metrics via an open-source viewer.

A foundational DevOps workshop was also conducted to support enablement and training, tailored to the transportation company’s context with a focus on Azure DevOps.

Tech Stack used:
  • Github Copilot
  • VSCode
  • Microsoft Teams (for meeting transcript)
  • Java – Spring Boot
  • NextJS

Architecture Diagram

Outcomes

The DevPulse engagement delivered measurable improvements across several dimensions:

  • Team Collaboration & Agile Planning: By consolidating metrics from disparate tools into a unified dashboard, DevPulse enabled cross-functional teams to align on priorities and track progress at the product, initiative, and board levels.
  • Project Planning & Quality Enablement: The integration of SonarQube and GitHub Copilot metrics allowed the business to monitor code quality and developer productivity, supporting more informed planning and continuous improvement.
  • Software Testing & Automation: The ingestion pipeline and dashboarding facilitated automated tracking of test coverage and deployment frequency, reducing manual overhead and surfacing bottlenecks.
  • Deployment Speed & Uptime: Although specific uptime metrics were not disclosed, the architecture supported real-time data flow and visualization, contributing to faster decision-making and deployment cycles.

Phase 2 – GHCP Deep-dive Trainings

Post the DevPulse POC engagement, addressing the low adoption rate of GitHub Copilot became crucial. PALO IT conducted three rounds of deep-dive trainings to better educate the company’s developers on the tool’s full potential. The agenda included:

  • An introduction and refresher of GitHub Copilot.
  • Unusual use cases of GitHub Copilot, such as bash scripting, Terraform policy compliance, and documentation generation.
  • Challenge-based scenarios customized for the company’s development team.

Feedback received from these PALO IT Academy trainings was overwhelmingly positive, with participants indicating high satisfaction levels 👍

Phase 3 – Gen-e2 GenAI Delivery

Context & Business Challenge

The company's internal IT department encountered a significant productivity plateau within their Agile teams, compounded by silos among business, design, testing, and development. They sought fresh, innovative solutions to unlock new productivity gains and enhance their software delivery processes. The goal was to validate that AI-enhanced software development could deliver tangible benefits, including increased velocity and reduced costs.

Solution

PALO IT introduced Gen-e2™, a responsible AI-first product engineering framework that integrates GitHub Copilot into the entire software delivery lifecycle. This approach reinvents the traditional software development lifecycle to be AI-driven, allowing for the generation of 95% of high-quality code, along with necessary architecture diagrams, documentation, tests, and infrastructure.

The process utilizes “Context Prompting,” enabling the AI to hold comprehensive context about the product being built. This facilitates more efficient prompting and reduces repetitive input from engineers.

AI Delivery Approach and Implementation

Implementation & Results

The implementation applied AI-enhanced techniques across a wide range of activities, including requirements gathering, user story writing, and feature development for frontend and backend. The pilot was planned over 14 weeks with 6 checkpoints, originally set for an 11-week development phase as shown above.

The results were astounding—what was initially projected as an 11-week development cycle was reduced to just 5 weeks, yielding:

  • Cost savings of 54%
  • Time efficiency improvements of 55%
  • Increased developer satisfaction and user interaction
  • Nearly 90% of team members indicated increased confidence in using AI tools by the end of the pilot

To give you a better snapshot—when the pilot began, only 340+ users were onboarded with GitHub Copilot. By the end, nearly 1,100 users were onboarded.

Forecasted vs. Actual Delivery Burn-up Chart

Conclusion

As they say, the proof is in the pudding. The client's success was real and measurable, best represented in the aforementioned metrics. The story goes on—following the successful pilot, the business expressed a strong interest in advancing its transformation journey. The client currently aims to explore broader implementation while ensuring the approach remains effective across different organizational contexts. This company's commitment to further training and scaling AI-driven product delivery positions them well for continued success in a market that's shown to be more and more competitive as time goes on. If this sounds like your cup of tea, get in touch with our team to learn more about the project, or kickoff your own AI transformation journey.