Using Generative AI in the development process: Where challenge meets opportunity
As part of Kainos’ commitment to developing cost-effective, safe and value-added Generative AI (GenAI) systems, we’re open-sourcing our AI Enhanced Platform (AEP). The AEP brings together a suite of technologies that enable fast user onboarding to Generative AI tooling via a secure platform.
The challenge and the opportunity
As organisations look to harness the opportunities presented by GenAI in their software development lifecycle (SDLC), we see the bleeding edge as the point where challenge meets innovation opportunity.
- How do we reduce adoption friction for our engineering teams?
- Provide the tools in a cost-effective way.
- Mitigate risks associated with accessing and using AI.
- Improve GenAI adoption visibility and track value.
- Centralise cost and access management for teams.
- Experiment quickly and share outcomes more easily.
Meeting the challenge
From the list of challenges we identified, we see the main issue being friction in the development process.
No matter how cost-effective or safe a tool is, if developers cannot use it effectively, the opportunity is lost. With usability at the forefront of our minds, we set out to identify the friction points:
- Engineering teams need a straight-forward onboarding experience.
- Tools should be non-disruptive to the development process.
- Experimentation and sharing across teams should be easy.
Considering the challenges around adoption, cost and access – the solution became clear. We need a platform that reduces complexity, helps us manage access, and provides clear metrics on usage.
Building the solution
Through an API, the AEP allows engineers to interact more easily with large language models in their project environments. The AEP stores pre-defined prompts, which enables your teams to share workflows across projects. To facilitate experimentation, you can send custom prompts to the AEP before storing them in the service. In-turn, your organisation is empowered to track the value provided by Generative AI across projects using service-generated metrics.
By leveraging Azure’s OpenAI Service, the AEP is able to deliver enterprise-grade privacy, security and compliance capabilities. From stopping prompt attacks to content filtering, we’ve ensured that the AEP is safe to use for enterprise projects. For a more in-depth technical view of the AEP, read here.
Realising the opportunity
The AEP is the basis we used to experiment with GenAI in our software development lifecycle. However, we want you to see immediate value from hosting the AEP in your organisation, so we’re also sharing the experiments we’re conducting with the AEP in our build systems.
The first version will include:
AI-driven pull request summary: Get a summary of your pull request using AI. Supported on: GitHub and Azure DevOps.
AI-driven pull request review: Get AI-driven insights on your pull request changes. Supported on GitHub and Azure DevOps.
AI-driven pull request review with board context: Get a review of your pull request and an analysis of its completion status based on the linked work. Supported on Azure DevOps and Azure Boards.
Get started today
If you want to improve your GenAI adoption, centralise cost and access management, and enable your engineers to experiment more quickly, follow these steps:
- Fork our open source repository.
- Follow the hosting guide.
- Start using our community tooling in your build systems – and get experimenting.
We’ll provide regular updates and new community tooling for the AEP as technology capabilities increase. If your organisation needs support in hosting our AEP, contact us for further information.