How Sparq Uses Generative AI

We started using generative AI tools at Sparq early in 2023. We chose the top toolset at that time to start evaluating and building out a rubric around it: what languages do they purport to support, IDEs, versioning, libraries, how well do they support those and started really doing a deep dive on it. We also rolled out to our customers an option for us to utilize generative AI on their projects.

Where We’re Seeing Success

Blended teams work really well with these tools. You have this mix of really experienced engineers and really new engineers and together they work through this really well. The feedback that we’re getting so far and what we’re seeing is 50% plus increase in productivity and efficiency for boilerplate or tedious code. Where you don’t see as much of a return is on business-specific or industry-specific logic, which makes sense. The model hasn’t been trained on as much of that, so you have to have the right processes in place, you have to have good code reviews, security scans and all the things that you would normally do as best practices, you have to still do them. This doesn’t negate the need for good process and if anything, it just encourages you to continue to do that.

Ensuring Readiness

Every CEO out there says we should be doing something in AI but you may not be ready from a maturity standpoint to actually do it; your data may not be ready, your systems may not be ready, so what do you need to get there? We’re also going into creating an AI readiness assessment for our clients. We’ll be asking questions like “Where are you in that? Where are you beyond generative AI? What are your goals, your drivers and your business needs?” Is it predictive analysis? Is it just better insights? Is it some kind of machine learning to create efficiencies? Where are you in the continuum?”