The Tuning Fork Problem: Why Most Brands Are Answering Questions Nobody’s Asking
On building AI systems and messaging that sense the audience's actual felt need, not their surface-level query.
Strike a tuning fork near a piano and the matching string hums back, untouched. Nothing hit it. It simply recognized itself. That's resonance — and it's the thing almost no brand messaging is built to produce.
Most brands are struck keys, not tuning forks. They answer the question a prospect typed, in the words the prospect typed it, and call that understanding. But the words in the search bar are rarely the words for the ache underneath them. Someone searching "brand strategist for AI startup" isn't shopping for a service category. They're trying to stop feeling like their company is invisible to the exact people it was built for.
AI systems inherit this same mistake, at scale. A model trained to match query to answer will happily serve a technically correct response to the wrong question — the surface one — and call it done. It has no mechanism for noticing the gap between what was typed and what was meant, unless something upstream taught it to look.
That something is what I call the tuning fork problem: the felt need sits one layer beneath the literal query, in a register most brand systems, and most AI training data, never reach. Closing that gap — for a human reader and for the model reading on their behalf — is the actual work of brand intelligence.