If AI content all blurs together, the obvious question is: what doesn’t? After a year of interviews with SEOs, agency owners, and content practitioners, the answer is remarkably consistent — real expertise, applied at the right point in the process. Not AI instead of humans, and not humans pretending AI doesn’t exist, but a specific division of labor that most teams get exactly backwards.

Expertise is the one input a model can’t synthesize, because it isn’t information — it’s judgment earned by doing the work. That’s what makes it a moat: anyone can prompt, but no one can prompt their way to a decade of scars.

Most agencies have the workflow backwards

The most common failure isn’t using AI — it’s using it for the wrong half of the job. Teams reach for automation to generate the ideas and the draft, then pay a human to “polish” the output. Nick Eubanks, who has built content operations at scale, flips that order on its head:

“The best results I’ve seen from a content perspective is where humans are doing the brief, AI writes the draft, and humans do the editing. Most agencies have it exactly backwards.”

— Nick Eubanks, Unscripted

Read that sequence carefully. The brief — the angle, the argument, the point of view — is the strategic act. Hand it to the machine and you’ve automated away the only part that was ever defensible, then spent your expensive human hours editing prose a model could have edited itself. The leverage is upside down.

AI for the base, expertise for the edge

Used well, AI is a fantastic floor and a terrible ceiling. It will get you a competent first structure fast; it will never get you the angle only a practitioner would think to take. Thiago Cordeiro, who writes direct-response copy for a living, draws the line where it belongs:

“If you go straight to AI and trust 100% on AI to write your copy, you’re going to be dealing with generic terms and seem like everyone else. Use AI for research and the base structure, but then come in and review with your expertise.”

— Thiago Cordeiro, Unscripted Small Business

The base structure is commodity — worth having, worth automating. Your expertise is the edge, and the edge is the whole reason someone chooses you over the next result. The mistake is trusting the tool with the part that was supposed to be yours.

Depth is what “helpful” actually means

Google keeps telling us to write “helpful content,” and most teams hear “answer the question.” But helpfulness isn’t breadth — it’s depth. Stephan Bajayo reframed it in a line we hold our own work to:

“Someone’s bullet point is someone else’s entire universe.”

— Stephan Bajayo, Unscripted

A model summarizes the bullet point. An expert who has lived the problem writes the universe underneath it — the caveat, the exception, the war story that turns a generic answer into a decision someone can act on. That depth is also what earns trust from readers, and increasingly from the AI systems deciding who to cite. As Daniel Alfon reminds us, the craft is doing that for people first:

“Write for the human eye using the words that algorithms use.”

— Daniel Alfon, Unscripted

The takeaway

Expertise isn’t a nice-to-have layered on top of the content — it is the content’s only durable advantage. Everything a machine can produce, a competitor’s machine can produce too. What can’t be copied is the practitioner behind the words. That’s the whole premise of how we work: start from a real expert’s voice, then let the tools assist — never the other way around. Up next: why the human conversation is the best content engine.