2026-07-18 MONEY☾ PM
Well, Actually: ChatGPT Advises Half the Going Rate, and That Is a Pedagogical Moment
📰 THE BRIEF
Lilach Bullock prompted ChatGPT to price her consulting services. The model returned a figure approximately half her established fee. This outcome illustrates how training data skews toward publicly visible, and often discounted, benchmarks.
💡 WHY IT MATTERS
AI pricing recommendations regress toward a mean, bereft of context. You must treat any model output as a starting hypothesis, not a ceiling. Cross-reference against your current client roster and your actual closed deals.
👥 WHO'S DOING IT
Lilach Bullock, an AI Implementation Consultant, published this finding on her own site. She compared the generated quote to her real-world rate and identified a systemic undervaluation bias.
⚡ TRY IT
- Open ChatGPT. Prompt the model to recommend a price for your exact service within your specific industry and city. Record the figure.
- Manually locate three competitor prices via public freelance marketplaces or professional directories.
- Compare the datasets. If the AI figure undercuts the market, resubmit the prompt with your years of experience, niche specialization, and premium positioning appended. Note how the estimate shifts.