Companies have used AI to generate code faster, but now everyone needs to shift back to a focus on code quality, especially in the face of increasing vulnerabilities. Doing this requires focusing on visibility, architecture, and engineering priorities.

In 2026, speed was the main metric: how fast code could reach production and how many features were shipped. 2027 is going to be the year that we focus on code quality because speed is now a commodity.
What's driving this change? AI is increasingly used to write software and find flaws in software, but not all code that was written before AI has been evaluated. There's going to be a large effort to go back and check the things that we thought were safe that may no longer be safe.
At the same time, the bad guys are using currently open-weight models that are as capable as some of the frontier models, but without the guardrails. With open-weight models and frontier models both scouring your code base, it is all the more likely that sloppy or bad code will be found and exploited.
When I say quality, what I mean is the code works as intended, is safe, and is simple and practical to defend. How fast it's released matters less than how defensible it actually is. There is still a need for speed, and patching is going to be the place where speed is still of the essence. For example, the PHP2Shell vulnerability was only hours old before it started being exploited.
Boards and CEOs are going to be asking, "Who is affected? Are we safe?" This is going to happen over and over again with different flaws that are found. Those answers are going to be hard to produce for most companies unless they actually have good patch management and a solid idea of what software actually exists and lives inside their boundaries.
Luckily, speed and quality are not a trade-off. It's not a zero-sum game. The safest approach should actually also be the fastest approach because it's the one that you have the most confidence in. That means governing the whole process from spec to production to runtime with policies that protect those assets.
Going forward, companies should be focused on:
1. Visibility: Know what's in the environment and how quickly you could tell leadership whether a new vulnerability affects your company or not.
2. Architecture: Restructure and rewrite your software and your systems to reduce what's exposed to the internet while keeping the same functionality, rather than just relying on scanners and putting anything out there that your developers feel like.
3. Engineering priorities: Product managers and engineering leaders need to shift from feature date KPIs to refactoring and revisiting architectural decisions, and making things work more efficiently with less code and less exposure. Users prefer an experience that is fast, works well, and is safe over one that has a fancy new gadget.


