The Client
United Rentals is the world's largest equipment rental company, operating a network of more than 1,500 locations across North America and Europe. Serving construction, industrial, and commercial customers, the company offers a broad range of equipment and tools supported by digital platforms and services that help customers manage projects more efficiently. United Rentals is publicly traded on the New York Stock Exchange.
The Situation
United Rentals partnered with Acquia partner, VML, to evaluate how AI-enhanced tooling could transform digital project delivery across the complete software development lifecycle. The engagement was designed to move beyond AI experimentation into practical, measurable adoption, understanding exactly where AI delivers value and where human oversight remains essential.
The engagement’s objectives were:
- Improve velocity and efficiency by achieving 25-30% effort reduction across project phases, including Define, Design, Develop, and Test
- Enhance quality and collaboration by identifying AI tools that amplify human expertise rather than replace it
- Enable organizational scalability by identifying tools suitable for broader adoption across development teams
- Capture actionable insights to inform future digital initiatives and establish best practices for AI integration
The Challenges
As United Rentals embarked on this evaluation, several significant challenges emerged:
- AI hallucination and quality risks: AI tools can generate content that appears authoritative but contains errors, meaning efficiency gains from AI generation were partially offset by increased review requirements. Thorough human oversight of all AI output was paramount to maintaining production quality.
- Trust deficit in automation: Particularly in testing and smoke testing, there was organizational hesitation to allow AI tools to perform tasks independently. UX issues not covered in the initial requirements were only uncovered through manual testing, reinforcing the need for human judgment in the process.
- Role boundary challenges: Experimentation with role shifting, such as non-technical staff generating code and developers creating test suites, revealed that preserving existing roles was necessary to ensure output quality. Platforms producing functioning prototypes delivered products of poor quality, unsuitable for production systems.
- Rapid AI evolution: During the evaluation period from August to October 2025, the AI landscape shifted dramatically, with new models, integrations, and tools launching continuously. This underscored the need for continuous evaluation rather than one-time tool selection.
- Integration and security considerations: Evaluating tools required balancing usability against security requirements, integration complexity with existing workflows, and total cost of ownership, particularly for enterprise deployment at scale.
The Solution
VML spearheaded a cross-functional team that evaluated AI tools using the Total Rental Cost Comparison Tool as its test vehicle—a production feature built on Drupal running on Acquia Cloud Platform. The team’s methodology was rigorous, as they performed identical tasks with multiple tools and compared results across every SDLC phase. The team tested ~15 different AI-enabled tools as part of this project.
The phased implementation included:
Define Phase
AI transformed ticket writing from a manual, time-consuming process into rapid, high-quality requirements generation. Planning and categorization also benefited from AI's ability to identify dependencies and logical groupings, accelerating a phase that typically demands significant human coordination.
Design Phase
Manual sketching and wireframes were replaced with functional AI-generated prototypes, enabling rapid exploration of varied solutions. These prototypes proved more effective for vetting UX flows and homing in on the right solution faster than traditional methods.
Develop Phase
AI excelled at architecture planning, revealing constraints that could otherwise be missed, and at backend development, rapidly producing standardized code when given existing patterns. Drupal on Acquia Cloud Platform provided the established patterns AI needed to generate contextually appropriate code that integrated seamlessly with production systems.
Test Phase
Documentation and test planning saw the strongest gains in this phase, with AI using interview-style questioning to ensure comprehensive coverage and identify testing gaps that might otherwise have been overlooked.
Throughout every phase, Drupal on Acquia Cloud Platform served as the stable, well-architected foundation that made AI acceleration possible. The established patterns, coding standards, and deployment infrastructure allowed AI tools to generate code that integrated seamlessly with production systems, demonstrating that a well-structured platform is a prerequisite for effective AI-assisted development.
The Results
The evaluation delivered measurable efficiency gains across every phase of the software development lifecycle. In the Define phase, the team achieved approximately 30% reduction in overall effort, with ticket writing showing the most dramatic improvement. The Design phase saw a 25% reduction in timeline while simultaneously increasing the scope of UX concepting. Both the Develop and Test phases achieved approximately 25% reductions in overall effort, with architecture planning and backend development driving gains in the former, and documentation and planning delivering the strongest results in the latter.
Productivity metrics reinforced these findings. The Developer Experience Index scored 77, placing the project between the P75 and P90 benchmarks. Time spent on new features reached 64.7%, also within the P75-P90 range, while PR Throughput came in at 3.5, at the P50 benchmark.
Across all phases, the team identified a set of sweet spots where AI delivered both high quality and high efficiency: UX concepting, development planning, unit testing, test planning, and writing new requirements all emerged as areas where AI integration produced the strongest combined outcomes.
The project's strategic value extended beyond the immediate metrics. The engagement produced comprehensive tool procurement recommendations for enterprise deployment, created the foundation for a living AI playbook for squad rollout and employee onboarding, and established measurement frameworks for ongoing AI impact assessment.
At a broader level, the project produced six key findings that will shape United Rentals' approach to AI going forward:
- AI delivers consistent 25-30% efficiency gains across most SDLC phases
- Human expertise remains essential; AI accelerates skilled professionals but does not replace domain knowledge
- Backend tasks see significantly better AI results than design-matching frontend work
- Early phases, including concepting, planning, and documentation benefit most, while execution phases show mixed results
- Continuous evaluation is necessary, given the rapid pace of AI evolution
- Role preservation ensures production quality, as AI augments existing expertise rather than enabling role substitution