05 AI · Tools · Standards · Self-directed

AI innovation in
content design

I built a custom AI agent for content design, created content standards for AI-generated experiences, and started AI office hours for the CD org — all before anyone asked me to. In a world where companies are asking "do we still need content designers?", I'm the one defining how AI writes.

3
Self-directed AI initiatives
0
Assignments required
1st
CD to run AI office hours
Custom
AI agent built for CD workflows
Org-wide
Standards adopted across CD org
The context

The question everyone
was afraid to ask.

In 2023, the conversation across the design industry shifted. Companies started asking whether AI would replace content designers. Some started cutting headcount. Others started quietly experimenting with AI-generated copy.

I didn't wait for the conversation to reach me. I went toward it.

I started exploring AI tools for content design workflows before it was a team priority, before it was assigned to me, and before most of my peers had formed an opinion on whether it was worth their time. I built a custom AI agent. I wrote content standards for AI-generated experiences. I started AI office hours for the CD org.

The answer to "do we still need content designers?" isn't defensive. It's this: content designers are the people who should be defining how AI writes.

The reframe
AI doesn't replace the judgment that makes content good. It amplifies it — or degrades it, if no one with that judgment is in the loop. I put myself in the loop before anyone asked.

"Ashlee has been ahead of the curve on AI adoption in content design. She's not just using the tools — she's defining how they should be used."

— Peer review, 2024
The work

Three initiatives.
All self-directed.

01
Custom AI agent for content design workflows
I built a custom AI agent trained on Meta's voice and tone guidelines, content standards, and wearables-specific patterns. The agent could generate first-draft strings, check copy against style guidelines, flag potential accessibility issues, and suggest terminology-compliant alternatives. It wasn't a generic LLM wrapper — it was a tool built around the specific judgment calls that content design requires. I documented the build process and shared it with the CD org as a replicable model.
02
Content standards for AI-generated experiences
When AI starts generating content that users interact with — notifications, suggestions, summaries, error messages — someone has to define what "good" looks like. I authored the strategic framework for how AI agents should communicate in wearable device experiences: when to be explicit about AI involvement, how to handle uncertainty, what tone is appropriate for different autonomy levels, and how to maintain user trust across the full spectrum from manual to autonomous. This framework was adopted in AI-first sprint planning.
03
AI office hours for the CD org
I started AI office hours for the content design org — a recurring session where CDs could bring questions about AI tools, share experiments, and develop shared norms for how AI fits into our workflows. The sessions covered prompt engineering for content design, evaluating AI-generated copy, maintaining voice and tone in AI-assisted workflows, and the ethical questions that come with AI-generated content at scale. I was the first CD in the org to run dedicated AI office hours.
The framework

Mapping the autonomy spectrum

The most consequential piece of this work was the agentic wearables content framework — a strategic document that mapped the full spectrum of AI agency on a wrist-worn device, from fully manual to fully autonomous.

The framework addressed questions that nobody in the industry had answered yet: When should an AI agent on a wrist act without asking? When should it suggest and wait? When should it be completely silent? How do you communicate uncertainty to a user who's mid-workout, driving, or in a meeting?

I wrote this framework before it was assigned to me, before there was a product requirement for it, and before most of my peers had thought through the implications. It was adopted in AI-first sprint planning as the canonical reference for how AI agents communicate on the device.

The spectrum
From fully manual (user initiates everything) to fully autonomous (AI acts without asking) — with clear content principles for every point on the spectrum, including when silence is the right answer.
Manual User-initiated, explicit confirmation required
Suggestive AI suggests, user decides — clear attribution
Proactive AI acts, then informs — undo available
Autonomous AI acts silently — trust established, stakes low
Why this matters

The content designer's role
in the AI era

The companies that are cutting content designers because of AI are making a category error. They're confusing "AI can generate text" with "AI has content judgment." It doesn't. AI can produce strings at scale. It cannot decide which strings are appropriate for a user who just got a health alert, or how to communicate uncertainty without eroding trust, or when silence is more respectful than a notification.

That judgment is what content designers bring. My job isn't to compete with AI — it's to be the person who decides how AI writes, what it says, and when it should say nothing at all.

I built these initiatives before they were assigned to me because I understood what was coming. The same instinct that led me to build the accessibility practice before EAA was a concern led me to build the AI framework before AI-first products were on the roadmap.

"In a world where companies are asking 'do we still need content designers?', Ashlee is the one defining how AI writes. That's not a defensive position — it's a leadership one."

— Peer review, 2024
Outcome

Defining the standard before
the standard existed

The agentic wearables content framework was adopted in AI-first sprint planning. The AI office hours became a recurring resource for the CD org. The content standards for AI-generated experiences became the reference document for teams building AI features on the device.

None of it was assigned. All of it was adopted.

What this proved
"The content designers who will thrive in the AI era aren't the ones who resist it or defer to it. They're the ones who define how it works — before anyone asks them to."
3
Self-directed AI initiatives
Adopted
In AI-first sprint planning
1st
CD to run AI office hours