Skip to content

    From hype to reality: Delivering value faster with AI at Cúram

    Published September 4, 2025 | 7 min read
    software developer using a monitor

    Every software leader feels the pressure: the demand to innovate faster, the challenge of modernizing complex systems, and the constant need to empower developers to do their best work. At Cúram, we were no different. We saw the hype around generative artificial intelligence (AI) and, like many, were optimistic but a little skeptical - could an AI-powered tool truly transform how we build enterprise-grade software for the public sector? 

    We decided to find out. The results were more transformative than we imagined. 

    This blog shares our story. It’s a look at the real, practical ways AI has delivered for us and the key lessons and actionable insights we’ve learned along the way. 

    AI in practice: software development use cases at Cúram 

    Our work with GitHub Copilot is a real-world example of how AI can deliver measurable value, fast. Copilot’s ability to assist in coding by suggesting patterns, completing arguments, and supporting the generation of test cases has drastically reduced manual lookups and tedious typing. This isn’t about replacing human creativity; it’s about augmenting it, freeing our developers to focus on the high-level architecture and logic that differentiates our solutions.

    The data speaks for itself. Since adopting Copilot as an optional tool for the team we’ve seen a dramatic increase in AI assisted Pull Requests (PRs), the method developers use to propose changes to the codebase. Starting at just 12% in the first month, the adoption rate climbed to 35% by the third and soared to an impressive 76% within six months, a level it has sustained ever since. This rapid, consistent integration demonstrates just how valuable AI has become in our daily workflows. 

    IMG-Curam-Web-Blog-AIGraphic-001

    Here are the key ways AI is augmenting our software development teams: 

    • Accelerated code modernization:  As part of modernizing our codebase, AI has been pivotal. Eoin Kelly, Principal Software Developer at Cúram, explains, “Copilot provides real-time assistance for complex questions and code generation, allowing us to modernize faster. For Cúram’s clients, this behind-the-scenes work means our solutions are more adaptable and efficient, which reduces long-term maintenance costs.”
    • Smarter, faster debugging:  Syntax corrections, code reviews, and identifying unreachable code are now significantly more efficient. Giuliana Giuffrida, Lead Software Developer, highlights the productivity boost: “It’s a game-changer when working with a language I’m less familiar with. I know the logic the code needs to follow, and I can ask Copilot for the correct syntax—for example, when writing a new Python script to analyze log files. It streamlines the process, making it faster to implement features, even in unfamiliar coding languages.” 
    • Enhanced code clarity and collaboration:  By helping developers annotate their work, AI improves team collaboration. Ramprasad Bhat, Principal Software Developer, shares, “Beyond code generation, we use AI to help rephrase sentences and generate concise summaries in comments and documentation. It’s invaluable for navigating complex codebases, enhancing clarity, and saving time.” 
    • More efficient and comprehensive testing:  Writing robust tests has become significantly faster. As Giuliana notes, “We recently migrated our tests to a new framework. Copilot provided a helpful first draft. While we still made corrections, and updates, the AI-driven insights made the entire process much more scalable.” 

    These AI-powered enhancements have become a daily practice at Cúram, driving continuous improvement that directly translates into developer productivity and satisfaction, and ultimately, client value. 

    Crucially, we learned that AI adoption is not just a technical challenge—it’s a human one. Initially, some of our developers were skeptical. They were concerned about what AI meant for their roles and whether it would actually help. This is where our “champions group” became essential stakeholders in the process. 

    They didn’t just promote the tool; they shared their honest experiences. They validated that while the AI isn’t perfect and often needs to be corrected, and always needs to be reviewed, its ability to accelerate tedious work was still valuable. This freed them up to focus on critical thinking, complex problem-solving, and decision-making. That honest, peer-to-peer validation was far more effective than any management directive and was the key to building trust across the team.  

    Delivering value faster with AI-powered automation tools 

    By leveraging AI, we deliver features and improvements to clients with greater speed. Time-consuming tasks, such as debugging and test writing, are streamlined, thereby accelerating our turnaround times. 

    For social services providers using Cúram solutions, this means quicker updates and enhanced system functionality. These efficiencies translate directly into better outcomes—both for caseworkers who use our systems every day for complex case management, and for the individuals navigating social programs to access their benefits. 

    “By leveraging AI, we deliver features and improvements to clients with greater speed. For social services providers using Cúram solutions, this means quicker updates and enhanced system functionality. These efficiencies translate directly into better outcomes—both for caseworkers who use our systems every day, and for the individuals navigating social programs to access their benefits.”

    -Mark Curtin

    Key lessons learned from our experience 

    The introduction of AI at Cúram was a deliberate, phased process. Here is the playbook we developed from our experience: 

    1. Start with intentionality.  AI isn’t a magic solution. Successful adoption demands clear project goals. We initially aimed to increase test coverage while reducing the time spent writing tests. Setting measurable objectives helps teams focus their effort and learning and ultimately integrate AI more effectively.
    2. Foster continuous learning.  AI tools and supporting AI models evolve rapidly. Maintaining a student’s mindset of curiosity and adaptability is essential to harness the full power of AI. To create this iterative journey at Cúram, we supported our teams with: 
      • Targeted training sessions on best practices and methodologies. 
      • Comprehensive guides and documentation. 
      • Collaborative spaces for sharing tips and success stories.
    3. Empower champions to drive adoption.  We launched a “champions group” of early adopters to test tools, provide feedback, and support their peers and team members. These champions were instrumental in refining our rollout and encouraging team-wide adoption.
    4. Frame AI as augmentation, not replacement.  AI tools are designed to complement human expertise. The goal is to let the machine handle the repetitive work so your team can focus on more creative and strategic tasks. 
    5. Prioritize security and trust.  Protecting sensitive information is non-negotiable. Our approach included: 
      • Continual reviews and application of our proven development practices to uphold quality. 
      • Using controlled environments:  For example, we use the paid, business version of GitHub Copilot to safeguard data and minimize risks. 
      • Thorough vetting: We partner with trusted providers like Microsoft and GitHub to maintain system integrity and protect our intellectual property (IP). 
    6. Let clear metrics guide your strategy.  Set clear goals and track your progress. Our data shows that our developers are now routinely leveraging AI to support their work. As a result, we see a reduction in the time it takes to resolve tickets while upholding our high standards for delivery – this demonstrates a significant return on investment and validates our strategy. Measurable objectives prove your success. 

    Ready to start your own AI journey? 

    For a deeper dive into artificial intelligence in the public sector, explore a practical six-step framework to support ethical and effective AI adoption in government services. It provides a structure to evaluate outcomes, risk, feasibility and more, for the adoption of AI in the public sector. Cúram applied this framework to identify software development as a low risk and high value use case to incorporate AI, in this case with GitHub Copilot. 

    Moving forward with artificial intelligence 

    At Cúram, we believe in embracing technology that empowers our team and enhances our clients’ success. GitHub Copilot is a prime example of how intelligent tools can elevate the development process, allowing us to build, innovate, and deliver solutions for social programs management with greater speed and precision. 

    For leaders in the public sector, particularly health and human services, AI is not a far-off concept but a present-day solution to streamline workflows and enhance service delivery. If your organization is considering this journey, remember our key takeaways: 

    • Start with a clear plan that begins by encouraging experimentation. 
    • Provide meaningful support and empower your team through education and the appointment of champions. 
    • Measure success with actionable metrics. 

    Ultimately, AI is more than a technological upgrade. It’s an opportunity to reimagine business processes and transform outcomes for the people we all serve. 

    How can AI be used in the public sector? 

    For deeper insights on how AI can be used in the public sector, specifically in health and human services, read the whitepaper from Cúram. The paper explores case studies on uses of AI that can help support government caseworkers deliver crucial services to constituents faster, help young people secure a better future, and help developers to efficiently support government modernization. 

    Read the whitepaper

    Ready for a consultation?

    Our team is ready to answer your questions. Let's make smarter health ecosystems, together.