Transcript — Episode 0007: The Last Read: WordPress 7 on a Live Editorial Workflow
Who’s speaking, and how. Christopher hosts in a professional clone of his own voice. The named guests — Frances, Walter, and Owen — are AI-synthesised personas built from ElevenLabs library voices, each one written by Christopher Ross. About the cast: https://thisismyurl.com/stances/i-disclose-every-use-of-ai/
Christopher: A few weeks ago, a team I know almost published a number that wasn’t true.
Christopher: It was a good number. Clean, specific, exactly the kind of stat that makes a paragraph land. An AI draft had written it in, and it read like it belonged.
Christopher: It didn’t ship. Somebody caught it at the last step before publish. Sixty seconds of reading saved a correction, an apology, and a small dent in a reputation that took years to build.
Christopher: Last time, we landed a rule for WordPress 7: trust the workflow features, and give anything that writes words a human read. Today I want to know if that rule survives a real publishing day. Because a rule is easy. A Tuesday is hard.
Christopher: Let me set the scene, because this is where the rule meets the floor. Picture a small team that publishes every day. Not a giant newsroom — three or four people, a lot of output, not a lot of slack.
Frances: That’s most of the teams I work with, honestly. And it’s exactly who WordPress 7 helps the most, because every saved minute actually shows up for them.
Christopher: Right. So they turn the features on. The interface changes are an instant win — nobody thinks twice. And they start leaning on the AI drafting for outlines, first passes, the scaffolding nobody enjoyed building by hand.
Walter: Which is the right way in. I’ve watched teams take on new tools for eighteen years, and the ones who do well treat a tool as a way to skip the part that was never the fun part anyway.
Frances: And day one, it’s lovely. Drafts come faster, the blank page stops being scary, everybody’s a little lighter.
Christopher: And then?
Frances: And then some ordinary Tuesday, somebody’s rushing, an AI-written line goes out with a fact nobody checked — and now you’re finding out where your process actually was, instead of where you assumed it was.
Walter: The tool didn’t fail there. It did exactly what it promised. It just found the gap that was already in the room.
Christopher: So here’s the real question, and it turns out it isn’t a technology question at all. Where does the human read actually happen?
Frances: And please don’t say “whenever someone has a minute.” Because I can tell you exactly when that is. Never. “When there’s time” never actually arrives — there’s always one more thing in front of it.
Christopher: Heh. Yeah. I’ve run that experiment on myself, and I’ve lost it every single time.
Walter: The honest version is a little uncomfortable. The faster you publish, the more drafts move through. And the more drafts move through —
Frances: — the more chances there are for one to slip past you. So the speed makes the read more important, not less. That’s the part people get backwards.
Christopher: It does feel backwards, though. You bring in a tool to go faster, and the takeaway is you need a checkpoint? Let me say it the way a skeptic would hear it — that sounds like taking the fast car and adding a speed bump.
Walter: But it isn’t a speed bump. A speed bump slows everyone, everywhere, forever. This is one gate, in one place, that you actually want there.
Owen: Can I bring the reader into the room for a second?
Christopher: Please, Owen.
Owen: Every one of those drafts becomes a page somebody lands on from a search, cold, trusting it. They don’t know an AI wrote the first version. They don’t know whether anyone read it after. They just believe it — because it’s on your site, with your name on it.
Owen: The human read is the one moment you get to stand exactly where that reader is standing.
Frances: Okay. That actually reframes it for me. It isn’t quality control. It’s the one moment somebody on the team is the reader instead of the writer.
Frances: So here’s where I’ve changed my mind. And I don’t do that out loud very often.
Christopher: Somebody mark the date.
Frances: Put the read at the moment of publish. Not earlier, where it’s theatre and nobody quite means it. Right at the gate. One named step, owned by one named person, every single time.
Frances: The publish button isn’t bureaucracy. It’s the last person who loves the reader.
Walter: That’s the whole thing, really. Not a committee. One person, one read, one moment.
Owen: And I’d make it small enough that nobody can talk themselves out of it. Three questions: is it true, is it ours, would we sign it. I put those on a card you can pin right by the publish button.
Christopher: Sixty seconds. And if all three are yes, you ship — fast, and with a clear conscience.
Christopher: So here’s where we land, and I find it genuinely encouraging.
Christopher: The team that decides where that gate goes, before they ever need it, ends up faster and safer than the team that never picked up the tool at all. You get the speed, and you keep the trust. Nobody made you choose.
Walter: The tool gives you the time. The gate keeps that time from quietly costing you something.
Frances: Decide the gate on a calm day. Not in the middle of the one where it suddenly matters.
Owen: And write it down, so it outlives the person who set it up.
Christopher: That number I mentioned at the top — the one that didn’t ship? It didn’t ship because somebody had already decided, weeks earlier, on a calm afternoon, that the last read was their job. The tool didn’t save them. A decision they’d already made did.
Christopher: Better interface. Careful AI. And a gate you chose on purpose.
Christopher: Thanks for spending this one with me.
Christopher: This is Sites I’ve Never Seen. Next time, we step back from the tools entirely, and ask what we’re actually chasing when we optimise a site — and what it costs when we chase too hard.
Christopher: Glad you were here.
End of episode.