A backend engineer accepts a Copilot suggestion that loads all customer records into memory and filters them with a nested loop to find matching orders. The code passes unit tests against the small test fixture and works fine in local development. During a code review, a senior engineer flags a concern about deploying this to production, where the customer table contains millions of rows. What is the most responsible way to handle this situation before merging?