What Happens When AI Becomes Your Junior Copywriter

In the absence of innovation, innovate 

I recently caught up with Tom Murphy, a former colleague and now Senior Content Designer at Webex. Mid-conversation, he drops a casual bomb: he built his own AI content assistant – no engineering help, no budget, just a clear problem and a sharper solution. It’s not a Webex-initiated product. It’s not proprietary tech. It’s just smart, scrappy and real. Exactly the kind of example that brings everything we’ve been saying into focus.

If you’ve never worked on a product team, you might not realize how much of the user experience is written or who’s writing it. That’s where content designers come in.

They don’t just write the words inside a product – they shape how it feels to use. They guide users through complex tasks. Anticipate confusion before it happens. Turn friction into clarity and clicks into confidence. Think button labels, help text, error messages, tool tips. But also onboarding flows, empty states, in-app education and the nuanced microcopy that makes a product feel smart, simple and human. 

And on Tom Murphy’s team, a handful of content designers support more than hundreds of designers. Which means they’re constantly getting pulled into last-minute requests, copy cleanup and string reviews. They weren’t designing experiences anymore, but rather, they were just fixing words.

Tom knew something had to give.

The DIY assistant that actually writes like you

Tom didn’t lobby for headcount. He didn’t wait for an AI budget. He got to work. In a single day, using a popular GenAI platform, he built an agentic content partner trained on the brand’s tone, style and accessibility guidelines – essentially giving every designer their own junior copywriter. 

Tom told me he uploaded the entire content standards PDF and trained the prompt to refer to it religiously. The outcome? Consistency at scale without slowing down the work.

He named it Write Like Webex, a tool that helps designers create better content before it ever hits the content team.

The agent:

• Guides designers through structured prompts to get on-brand content in real time
• Flags content that doesn’t align with voice and tone and explains why
• Supports non-native English speakers and junior designers with accessible feedback
• Reduces reliance on lorem ipsum and guesswork

“We were overbooked. Flooded with requests. We needed to check grammar, style, consistency. And that’s not the purpose of our role.”

Tom Murphy, Senior Content Designer, Webex

Not replacement, enhancement

This isn’t about replacing content designers. It’s about letting them work like content designers again. With AI handling the repetitive, tactical requests, the real humans can focus on the strategic stuff:

• Designing content hierarchies
• Coaching teams on how to prompt effectively
• Shaping the full product narrative
• Driving clarity through structure and UX, not just sentences

Tom’s team is also helping product managers get better at prompting, ensuring the copy that gets approved meets standards before it hits production. It doesn’t just assist with writing, but improves the thinking that leads to better product experiences.

Tips for building your own AI content partner

Want to build your own “Write Like Us” engine? Tom has a few pointers:

• Use AI to elevate humans. Don’t just plug in better copy. Give content designers more time for strategy, hierarchy and systems thinking, AKA the stuff machines can’t do.
• Start scrappy. Tom built his prototype in a day using natural language with zero engineers and zero budget.
• Build a knowledge base. Train the LLM using your real tone, style and accessibility standards. Make sure it always refers back to them.
• Structure your prompts. Use step-by-step Q&As, character count rules and content goals to generate content that fits the design. Not the other way around.
• Build trust. The AI model doesn’t just rewrite. It gives rationales. That builds trust with users and reinforces the brand voice.
• Use AI to elevate humans. Don’t just plug in better copy. Give content designers more time for strategy, hierarchy and systems thinking, AKA the stuff machines can’t do.