If you're using AI to develop market strategy, this little parlor trick will change how you think about your prompts. I'll give you a ChatGPT prompt that in no way details what AI should create, yet before you run it I can tell you with surprising specificity what the result will look like.
Here's the prompt:
𝘎𝘦𝘯𝘦𝘳𝘢𝘵𝘦 𝘢𝘯 𝘪𝘮𝘢𝘨𝘦 𝘵𝘩𝘢𝘵 𝘭𝘰𝘰𝘬𝘴 𝘭𝘪𝘬𝘦 𝘢𝘯 𝘪𝘗𝘩𝘰𝘯𝘦 6 𝘧𝘭𝘢𝘴𝘩 𝘱𝘩𝘰𝘵𝘰, 𝘵𝘩𝘢𝘵 𝘭𝘰𝘰𝘬𝘴 𝘤𝘩𝘢𝘰𝘵𝘪𝘤 𝘢𝘯𝘥 𝘶𝘯𝘤𝘢𝘯𝘯𝘺.
Run this on ChatGPT and I'll tell you what your image will look like. (You can do this on Gemini, but its training will render slightly different results.)
𝐒𝐏𝐎𝐈𝐋𝐄𝐑: Your generated image will look like the dark aftermath of a raging house party with creepy dolls and other odd characters looking into the camera amid a pile of trash.
Why does everyone get a similar result that doesn't seem predictable from the prompt? Because many of the words we use that seem unique are actually vectors for AI to zero-in on very common ideas:
• The iPhone 6, the best selling iPhone ever, filled social media with "improved" flash snapshots, featuring harsh lighting and deep shadows.
• The word "chaos" maps to messiness, clutter and disarray.
• The word "uncanny" maps to concepts related to "uncanny valley", which describes animated characters that appear uncomfortably human.
Three words that seem like creative direction are really coordinates for an LLM to lock in on the same idea.
There are two critical mistakes illustrated here that we also make using AI for strategy:
• We think AI digs in to all the details we feed it, but it's mostly looking for what it already recognizes.
• We think AI is delivering personalized responses, when we're doing the personalizing, similar to the way people read horoscopes.
It's hard to see this from inside your own conversations with AI. The output feels tailored because you wrote the input, and you rarely have the chance to compare with others. The problem is that AI regresses strategic thinking toward the mean.
The same happens when companies turn to AI to generate market strategy.
General competitive qualities are fed into prompts, touting the company's innovation, quality of service, technical expertise, customer focus, security and ROI, mirroring close competitors taking the same shortcuts and using the same playbook with AI.
Human consultants produce generic briefs too. But they charge enough that clients push back, demand revisions, and slog through the hard work to find differentiation. AI feels free and the result looks professional, so fewer people push back on the output. It's good enough.
It's hard work to generate a market strategy, which makes it a juicy target for using AI. But just because your result 𝘧𝘦𝘦𝘭𝘴 unique, doesn't mean it is.
AI won't tell you when you're falling into this trap. You won't know until the market shows your strategy is wrong. The hard work has to happen before the prompt or it won't happen at all.