How Lekko uses AI the smarter way

“Lekko has a huge advantage over SaaS tools to incorporate AI into coding.”

Every tech company now touts an AI angle to their story and their product pipeline. It’s a requirement to get investors to even look at a startup’s deck. For established firms, it assuages shareholders that their shares won’t be rendered worthless by competitors who embrace AI first and more fully. But for Lekko, AI was foundational before ChatGPT arrived, and we use it more wisely than we’ve seen elsewhere.

Lekko knows your code better than any SaaS

I founded Lekko with AI for dynamic configuration as one of its pillars. AI as it already exists, and as it is already seen to be evolving, has huge potential to remake software development. That’s true from the ground-floor basics of feature flagging to executive suite decisions on direction.

At Lekko we work internally with the presumption that AI will be writing more and more of our own code — and everyone else’s — under the guidance of skilled engineers. Lekko has a huge advantage over SaaS tools to incorporate AI into coding, since it lives in your code (privately and securely) rather than sitting remote to take API calls. Just as Copilot has read and memorized GitHub, Lekko will know how your own code works, This will give Lekko’s AI components far greater insight into how to develop and manage your own products. Let’s look at what exactly it will be able to do.

Determine the risks of a rollout

AI has right-now applicability to software development. Lekko, with direct knowledge of your code and confidential historical data, will be able to:

  • Spot untested code
  • Calculate the impact of change, i.e. “this change will affect 82% of your enterprise account holders.”
  • Learn what types of changes were risky, and recognize future risks better than engineers’ memories or checklists

Undertand the business impact of a configuration change

With access to statistics, sales records and user behavior, Lekko will be able to:

  • Estimate how a change to configuration or configuration options will affect revenue
  • Know in advance which customers are likely to be affected, and their levels of engagement
  • In all cases, be smarter than the smartest engineers at forecasting what will happen when a feature is flipped on or off

Connect and collaborate with other AI tools

Further down the pipeline are more strategic interconnections between Lekko and other AI-enabled tools to:

  • Give product managers an interface to understand the effects of adjusting configuration options
  • Collaborate with AI code-writing tools to minimize risk and enhance understanding of potential options
  • Simplify the use of and dependencies upon cascading chains of APIs

We seized the opportunity to start building Lekko in tandem with the nascent — but fast approaching — future of AI in software development. My view from here of that future is at once specific and fluxional. Some things are certain, some are maybe, and some none of us have yet thought of. For a programmer, there couldn’t be a more exciting time.