A CRO consultant told me our agent wasn't going to take his job. I don't think he's right.

"You built a good product, but it's not going to take my job."

A CRO consultant said that to me on a call, looking at an agent that does a good chunk of what he does. I nodded at the time. I don't think he's right.

We can already do a lot of what CRO consultants do. Not the judgment. The volume.

Almost everyone we talk to has the same problem. Hundreds of pages. The product ships every couple weeks. The market keeps moving. The ads keep changing. Something on the site is out of date the second you stop looking at it. And it's nobody's job to fix that, because no single stale page is ever urgent.

So the work gets rationed. A consultant does the five pages that matter most this quarter. Everything else quietly rots.

That's what we're building for. Keep the whole site current while the product and the market move under it. Match the page to the ad that drove the click. Make the new pages nobody has time to make.

The human part moves up. Supervising instead of producing. Setting direction. Making the call from experience that no agent would think to make, because it isn't in the data yet.

He keeps his job. He just stops rewriting headlines for a living.

To be concrete about the split.

The agent takes the volume. Read every page against what the product actually does this month. Match the page to the ad that drove the click. Write the variants, ship them, watch the results. Generate the pages nobody has time to make.

The person takes the parts that need a memory longer than the data. Which bets are worth making. What a customer meant rather than what they said. When the number is right and the conclusion is still wrong.

That's a better job than the one most CRO work is today.