How to clean up the tech debt your AI is creating | On-demand Webinar | Harness Resources

Tech was already facing a zombie (service) apocalypse. Now AI-driven development is reshaping the traditional distribution of tech debt. Abandonware is an industry epidemic, with 93% of open-source codebases containing components with no development activity in the last two years. While AI reduces minor, localized design-related issues, it causes a sharp increase in requirements debt and testing debt. Add to this, when AI is aggressively adopted without full developer understanding, confidence, or verification, research has found that AI creates a false sense of velocity among engineers. You’re actually going much slower than you think. Now AI agents could be making decisions on your behalf as to what should stay on and what should be switched off, often without any explanation. Technical debt created by agents in production is costing more than just tokens. It’s risking your security, your reliability, your reputation, and your whole business. In this expert panel for engineering leaders, we offer strategies for how to clean up your immense technical debt, so you can finally unlock AI speed at scale. Key takeaways: How to set the guardrails to avoid debt and enable AI speed. How to set the right context for your agents to minimize repeat code and reduce batch size. How feature flags, state-saving agents, automated debt detection, and other techniques can reduce your debt.