I translate tech debt into the language the audience already uses: money, speed, and risk. A non-technical stakeholder doesn't care about coupling or test coverage; they care that features are getting slower and more expensive to ship. So I frame debt as a financial and delivery problem, not a code problem.
