In this use case, we discuss how Capco has collaborated with Canadian payments fintech leader Interac Corp. to leverage GenAI to uplift the developer’s experience while preserving the confidentiality of their code and development processes.
It is predicted that in 2025 more than 80% of systems integrators and software providers will be using GenAI tools across multiple phases of the software development lifecycle, up from 10% in 2023.1 By 2027, it is expected that GenAI will augment 30% of all knowledge workers’ tasks.2
The use of Large Language Models (LLMs) offers unparalleled opportunities to enhance productivity and innovation. With the priority of protecting intellectual property (IP), Capco collaborated with our client Interac Corp. (Interac) to develop a private LLM-driven GenAI platform. This initiative ensures that Interac’s valuable IP remains securely in-house while still reaping the benefits of sophisticated AI technologies.
Jeff Henderson, Chief Technology Officer, Group Head, Technology & Operations Chief Technology Office at Interac Corp., “Interac is on its transformation journey to drive digital prosperity for Canadians.”
Kreshnik Mati, Head of Enabling Technology, Chief Technology Office at Interac Corp., “Building a GenAI platform that provides controls and guardrails is the first foundation step to a GenAI solution.”
WHY NOW?
In today’s digital era, ensuring data privacy and IP security has never been more important. Relying on external LLMs can pose significant risks, however, and companies worldwide are now confronted with the challenge of maintaining control over their proprietary information.
By developing a private LLM solution, our client aims to harness the power of AI without compromising the confidentiality of their code and development processes. The need for a private solution is further driven by the client’s commitment to fostering an innovation-focused environment where security is non-negotiable.
HOW WE DID IT
Our development strategy prioritized enhancing existing systems at Interac by introducing a suite of specialized plugins for a range of Integrated Development Environments (IDEs). These plugins were thoughtfully designed to integrate seamlessly, allowing developers to utilize multiple private LLMs from within their accustomed tools.
At the heart of our solution is the middle tier, which serves a dual purpose. Firstly, it provides the controls and guardrails and acts as an Enterprise GenAI platform providing security management, access management, persona management. It ensures that only authorized users have access to the LLM, thus safeguarding the intellectual property within the client's secure infrastructure.
Secondly, this middle tier incorporates comprehensive logging mechanisms. Every interaction with the LLM is meticulously recorded, allowing for an in-depth analysis of usage patterns and outputs. This logging capability provides valuable insights, enabling continuous improvement of the LLM tools. It also ensures transparency and accountability, as the logs can be reviewed to understand how the assistant is utilized and where further enhancements might be needed.
Overall, this design delivered a solution that not only facilitates developers' productivity but also maintains high security and provided a framework for ongoing optimization –without requiring extensive changes to their existing workflows.
WHAT WE DELIVERED
Security Enrichment:
- Developed an internal system that securely manages data exchanges, effectively eliminating the IP exposure risks associated with external LLMs.
- Implemented robust security protocols within our private LLM coding assistant to ensure all processes remain confidential.
Integration and Customization:
- Designed and integrated custom plugins tailored to fit seamlessly within the client's existing development environments.
- Ensured each plugin maximized usability and efficiency by aligning with established workflows and developer needs.
Scalable Framework:
- Introduced continuous tracking and validation mechanisms that serve as a comprehensive foundation for ongoing improvements and scalability.
- Established methods for collecting actionable insights that drive future enhancements.
Developer Experience Uplift:
- Created a more intuitive coding environment by enhancing development processes, which increased productivity without sacrificing security.
- Fostered an innovative development culture through the strategic empowerment of developers using sophisticated tools.
- Based on a benchmark test, the productivity of the developers at Interac can be increased by 50-80% on creating unit test cases, code review and code explanation.
CONCLUSION
As technological landscapes continue to evolve, so do measures for safeguarding intellectual assets.
Serena Chan, Partner, Head of Innovation & Design at Capco: “By embedding GenAI into core business processes, organisations can automate routine tasks, increase operational efficiencies, reduce cycle times, and derive actionable insights.”
This Interac project not only addressed immediate IP protection concerns, but also laid the groundwork for sustainable innovation. Looking ahead, we envision the ongoing refinement of our private LLM assistant, driving our client's goals for secure and efficient software development.
References
1 Emerging Tech: Primary Impact of Generative AI on Business Use Cases 6 September 2023- ID G00796157- By Annette Jump, Anthony Bradley, Eric Goodness, Radu Miclaus
2 Emerging Tech: Primary Impact of Generative AI on Business Use Cases 6 September 2023- ID G00796157- 26 min read By Annette Jump, Anthony Bradley, Eric Goodness, Radu Miclaus
14 March 2025
Authors: Serena Chan, Lucas Alba, Jimmy Kong