Managing "Shadow AI" at your agency

Most agency owners probably haven't thought much about shadow AI (the name for when employees use their personal ChatGPT, Claude, or Gemini accounts to do work). In fact, owners may be doing it as well. In this episode, Chip and Gini walk through what the risks are and how to respond without overreacting. The instinct to crack down is understandable but wrong. Employees are going to use their own tools regardless, often because personal accounts are better trained or more accessible than whatever the company has set up. The goal should be education, not elimination. Most employees don't know that personal accounts default to feeding data into training sets, or that a single toggle can turn that off. That one fix alone is worth a conversation with your team. Vibe coding and a plain-language AI policy get discussed, in addition to educating your team. Gini's team runs weekly micro-learning sessions to help people use AI as a thinking partner, not just a drafting tool. Both Chip and Gini advise that owners and employees who aren't using AI meaningfully within the next year or two are putting their careers and businesses at risk.