Accelerating Development at Scale: DXC London Market's Journey with Amazon Q Developer and Kiro

In the heart of London's insurance market, one of the industry's most ambitious digital transformation projects is underway. With hundreds of architects, engineers, developers, and testers working across multiple workstreams, DXC [1] for the London Market [2] (LM) is pioneering new approaches to accelerate development while maintaining the highest quality standards. Their journey with Amazon Q Developer  offers valuable insights for organisations contemplating similar transformations, particularly in how to move beyond surface-level adoption to achieve meaningful organisational change. As AWS continues to evolve its developer tooling—most recently with Kiro —organisations like DXC LM are well-positioned to leverage these advances.

The scale of LM’s transformation project presents unique challenges. With hundreds of architectures needing to integrate seamlessly, and teams working across multiple interconnected workstreams, the initial instinct was to focus purely on developer velocity. However, the leadership team, including Peter O'Hanlon, quickly recognised a more nuanced reality: "On a project of this size and complexity, developer velocity alone isn't where you ultimately need to land," Peter explains. "We needed everyone—developers, testers, architects—to move faster, but more importantly, to produce work that wouldn't fail in later testing phases."

This insight crystallized into a core principle that would guide their entire transformation: better quality, faster.

LM’s transformation journey was guided by AWS's Enterprise Adoption framework for Amazon Q, providing a structured approach to implementing Q Developer at scale. This framework proved crucial in moving beyond surface-level implementation to achieve meaningful organisational change, offering proven methodologies for 1/ Establishing clear success metrics, 2/ Creating effective feedback loops, and 3/ Developing sustainable adoption patterns.

LM’s approach to measuring and driving success evolved significantly throughout the year. They introduced adoption metrics (‘Level 1’), tracking basic usage patterns and interaction rates; while these metrics provided visibility into tool adoption, they proved insufficient in isolation. The team then focused on direct business impacts (‘Level 2’) such as velocity and defect rates, however the team recognised that measuring impact required a more nuanced approach.

The breakthrough came with the introduction of a comprehensive 10-week Developer Experience (DevX) study (‘Level 3’). This analysis tracked sentiment, time savings, challenges, and wins across the development community.

Most significantly, correlating all three levels of metrics revealed powerful insights: teams with the highest Q Developer usage, produced higher quality code, released faster, and showed both the greatest time savings with highest developer satisfaction levels.

The human aspect of the transformation proved as crucial as the technical implementation. The team identified distinct adoption patterns among their staff:

Surprisingly, junior developers showed initial hesitation despite their presumed comfort with new technology. "We were wildly optimistic that people would naturally embrace generative AI," Peter reflects. "We were wrong. Junior developers, being less established in their careers, were actually more cautious about adopting new tools that might affect their development journey."

As the adoption matured, the team discovered Q Developer's value extended far beyond traditional development tasks. Architects began using it to validate architectural decisions and review API integrations. The Architecture Review Board (ARB) incorporated Q Developer into their workflow for testing API definitions and ensuring compliance with security standards.

Testers found new innovative ways to leverage Q Developer across their testing lifecycle. Beyond basic test creation, they used the tool to generate comprehensive test scenarios that might otherwise have been overlooked. When analysing coverage, Q Developer helped identify gaps in test suites and suggested additional test cases. The tool proved particularly valuable for debugging complex integration issues, where its ability to analyse patterns across multiple systems accelerated problem resolution.

"What started as a developer productivity tool evolved into something that transformed our entire delivery pipeline," explains Peter. "From architecture validation to testing, from security reviews to documentation - Q Developer became integral to every aspect of our software delivery lifecycle."

The team's measurement approach evolved into a sophisticated framework that captured both quantitative and qualitative impacts:

Level 1 - Adoption Metrics; while basic usage statistics proved insufficient alone, they provided valuable context when combined with other metrics. Using the Q Developer User Activity Dashboard , the team tracked:

Level 2 - Business Impact Metrics; these metrics demonstrated concrete value to stakeholders, captured via custom method using Jira & Git:

Level 3 - Developer Experience (DevX); the 10-week survey-based study revealed:

A critical learning emerged early: self-reporting proved unreliable. By week three of the DevX study, the team stopped relying on developers' self-assessments of their Q Developer usage. "People are naturally optimistic," Peter notes. "When we asked if they were using the tool, we'd hear 'Yes!' enthusiastically, but the actual usage data told a different story." This led to a data-driven approach using objective metrics from Q Developer's operational dashboard, Git commit analysis, and Jira tracking—removing emotion and subjectivity from the measurement process.

The breakthrough in measurement came from correlating adoption metrics across the three levels; this correlation revealed the powerful insight that teams with the highest Q Developer adoption consistently delivered higher quality code, faster release cycles, and greater developer satisfaction—proving the business value beyond simple usage statistics.

The team's approach evolved from purely motivational to balanced accountability. Initial efforts focused on encouragement: sharing success stories, providing dedicated experimentation time, and creating platforms for teams to showcase innovative uses of Q Developer. However, this "softly softly" approach plateaued at around 50% adoption.

To further increase adoption, Leadership played a crucial role by introducing clear expectations and integrating Q Developer into standard AI-first workflows. Performance metrics began incorporating AI-assisted development practices, and team leaders were given specific adoption targets.

Team leaders implemented a blended scoring system that provided objective measurement across different roles. Each team member had a target of 500 points per week—not a rigid weekly requirement, but a trend-based indicator. This score blended multiple factors based on role: for developers, it included lines of code accepted and chat interactions; for architects, it weighted API validations and architectural reviews. This role-specific approach ensured fair measurement while maintaining accountability across the entire delivery organisation.

This combination of support and accountability proved essential in moving beyond early adopters to widespread organisational adoption.

LM’s future strategy focuses on three interconnected areas. First, they're working to standardise their measurement approach across all project types, moving beyond individual team metrics to create a comprehensive view of organisational improvement. This includes developing automated metric collection and detailed dashboards that provide real-time insights into the impact of AI-assisted development.

The second focus area involves expanding their successful adoption model beyond the initial teams. By standardising their adoption processes and creating centres of excellence, they're building a scalable model for rolling out Q Developer and Kiro across the broader Insurance organisation. This expansion is being guided by the lessons learned and best practices developed during their initial implementation.

Their AI strategy continues to evolve as they discover new use cases and opportunities for enhancement. The team is particularly focused on optimising the developer experience while pushing the boundaries of what's possible with AI-assisted development. This includes exploring new integration possibilities and finding ways to further streamline their development pipeline.

For organisations considering similar transformations, DXC London Market's experience highlights several critical success factors.

1/ First and foremost is the importance of comprehensive measurement. Organisations need to move beyond simple adoption metrics to understand the true impact of AI-assisted development on their operations. This means establishing clear baseline metrics before implementation and tracking progress across multiple dimensions.

2/ Cultural transformation proved equally important as technical implementation. Organisations must recognise that resistance to change is natural and plan accordingly. Success requires building a strong pioneer program while acknowledging that different teams and individuals will adopt new technologies at different rates.

3/ The technical implementation itself needs careful planning and execution. While robust infrastructure is essential, success depends more on starting with clear use cases and building on initial successes. Organisations should allow room for experimentation while maintaining focus on their core objectives.

4/ Perhaps most crucial is leadership alignment. Executive sponsorship must go beyond simple approval to active engagement and advocacy. Leaders need to align on core principles and maintain consistent messaging about the importance of AI-assisted development to the organisation's future. This alignment creates the foundation for accountability and sustained adoption across the organisation.

Ready to explore how Amazon Q Developer and Kiro can transform your development organization? Contact your AWS Account Team or Sales Support  to learn more about implementing AI-assisted development at scale.

[1] DXC Technology is a global IT services leader serving nearly 6,000 clients across 70 countries. As a premier AWS partner with over 10,000 AWS-accredited professionals, DXC has a significant presence in the London Market insurance sector, where the majority of their global insurance client base operates. DXC provides specialized solutions for London Market operations, covering commercial and specialty lines including property, casualty, marine, aviation, cyber, and emerging risks.

[2] London Market is the world's largest specialist insurance and reinsurance marketplace, accounting for 2% of the UK's GDP. The market consists of Lloyd's of London (with $57 billion in gross written premium) and the London Company Market ($46 billion), focusing primarily on non-life insurance and reinsurance, particularly marine, aviation, and high-exposure business