Where AI belongs in a copy workflow and where it doesn't

I'm not an AI skeptic. I build the AI tools and automations my team at Adoreal uses every day. But I've also read a lot of AI-written copy, and most of it has the same problem: It's fine. Fluent, grammatical and completely forgettable.

Fine doesn't win anything. So the real question is which parts of copy to use it for.

Where AI earns its keep

AI is great at the work around writing, the stuff that eats hours and doesn't need taste:

• Summarizing research, transcripts and long briefs
• Generating 20 rough headline directions so I can react to them
• Producing first passes on high-volume, low-stakes copy that a human will edit
• Checking drafts against a style guide or a list of banned claims
• Turning one finished piece into the formats each channel needs

That last one alone saves my team a ridiculous amount of time. A single strong article can become social posts, an email, a short video script and a sales one-pager. AI handles the first draft of each format. A writer makes sure every version still sounds like us.

Where it doesn't belong

That includes the big idea, the positioning, the homepage line that has to make someone feel something, an apology to an upset customer and any claim legal will read three times.

These need judgment about people, context and risk. AI can give you options. It can't tell you which one is true for your brand or which one your reader will believe.

It also can't sit in the room. Great positioning comes from interviews, sales calls, customer complaints and the odd comment a founder makes over lunch. AI only knows what you give it, and the most important insights rarely show up in a brief.

“Fine doesn't win anything. AI writes fine.”

The real skill is editing AI

The writers who get the most out of AI are the ones who were already great editors. They know what good looks like, so they can spot the generic bits, the confident nonsense and the phrases every AI tool seems to love.

Someone who can't edit will publish whatever comes out. That's how brands end up sounding like everyone else.

When I review AI-assisted drafts, I look for the same things I'd flag in any draft, plus a few extras: vague claims with no proof, tidy three-part lists that say nothing and conclusions that restate the introduction. Those patterns are easy to miss when the sentences are smooth.

Feed it better inputs

Most disappointing AI output comes from thin inputs. A one-line prompt gets you a one-line idea, stretched to fill a page.

The best results I've seen come from giving the tool real material: the brand voice guide, approved messaging, a strong example, customer language from reviews or calls and a clear description of the reader. Better inputs don't make AI creative. They make it useful.

Leaders set the rules

At a senior level, the job is deciding how your whole team uses AI: which tasks to hand off, what the quality bar is and where a human always signs off. That's a leadership decision, not a tool choice.

It also means being honest with your team about what changes. Some tasks will disappear. Others will grow. The writers who thrive are the ones who move toward strategy, editing and judgment, and a good leader helps them get there.

Measure what it saves

If AI is part of your workflow, track what it changes. At Adoreal, our automations cut a site search engine optimization audit from about four hours to 15 minutes. That kind of number makes the case for the tools and shows where the saved time goes.

The goal is more time for the work that needs a human, not more content for its own sake.

Keep your writers sharp

There's a quiet risk in leaning too hard on AI: Writers stop practicing the skills that make them valuable. If every first draft comes from a tool, it gets harder to write a strong one from scratch when it matters.

I make sure my team still writes the important things themselves. The homepage line, the campaign idea and the tricky customer message start with a person. AI can help afterward with options and checks.

That balance keeps the craft alive on the team, and it keeps the quality bar high. A writer who can produce great copy on their own is also the one who can tell when AI output isn't good enough.

The short version

Let AI handle the volume and the busywork. Keep humans on the ideas, the voice and anything with real stakes. Give the tools strong inputs, set clear rules for your team and measure the time you save. And invest in editing, because it's the skill that makes AI useful instead of just fast.