It’s rare to find a content design leader who has been trained to navigate AI disruption. For most, it’s been a baptism of fire. Atlassian’s Libby Varcoe and Arin Black share a framework for building trust, developing skills, and helping content designers shape what comes next — grounded in their two-year journey on the front line.

As AI reshapes how content gets created, content designers are facing a fundamental question:

If content can be produced at the speed of a prompt, what is the value of content design?

At Atlassian, we’ve been working through that question for the past two years. It has been messy, uncertain, and at times deeply uncomfortable. But it has also created an opportunity to move our craft upstream — to give content designers more agency, authority, and influence over the systems and experiences they help shape.

We shared some of this journey at Button 2026. Our experience is specific to Atlassian, but we suspect the underlying challenges are widely shared for teams of all sizes. Here are four principles that can help:

1. Shift the identity of the craft.
2. Support meaningful skill evolution.
3. Build trust through focused proofs of concept.
4. Create the psychological safety people need to change.

1. Shift the identity of the craft

In 2024, the conversation about generative AI was becoming impossible to ignore. New headlines arrived every week, many suggesting that content production would soon be automated — and that content designers might become obsolete.

Beneath the clickbait was a real shift. AI meant that content could be drafted, adapted, and distributed much faster. We knew we needed to articulate the value of content design in a way that made sense in this new environment.

The first step was changing the narrative around our work.

Instead of talking primarily about the local surfaces we supported — error messages, labels, modals, and other interface content — we began talking about designing for the system.

That meant positioning content designers as people who could:

• Design the structures and infra that produce content at scale
• Shape how AI systems behave and communicate
• Define terminology, concepts, and boundaries
• Create evaluation frameworks for quality
• Connect content decisions across products and channels
• Help make AI-generated experiences useful, understandable, and trustworthy

This shift reflected what content designers had been asking for over many years: the opportunity to work more upstream, influence the bigger picture, and contribute to the conditions that make good experiences possible.

We began with a small design sprint led by our Lead Content Designer, Sabrina Wingfield. At the time, access to AI tools was limited, so some of the work was speculative. The team explored ways AI could augment Content Design, including synthesising comments, drafting SEO descriptions, and aggregating research during discovery. While these early ideas were fun to explore, it was the mindset shift that mattered more. Our team quickly realised that we could play a role in how we use AI to shape the future of the craft, rather than it happening to us.

We were also deliberate about keeping content designers at the center of the story. We tried to avoid describing ourselves as passive reviewers or “humans in the loop.” Content designers could be the people who create and architect the loop — defining what the system should do, how it should behave, and how its quality should be judged.

This evolution eventually informed new role profiles and specialisations, including Content Technology and Model Design which we are really excited about.

Content Technology recognises the role content designers play in designing the infrastructure and systems that produce high-quality, AI-native experiences. It includes structured content, models, tools, pipelines, metadata, and the technical foundations that allow content to work across products and channels.

Model Design focuses on how AI systems behave in interaction: their instructions, decisions, conversational patterns, boundaries, and quality over time.

These specialisations complement, not replace, fundamental content design skills. Content strategy, systems thinking, information architecture, research, and quality measurement remain essential. In fact, as AI and automation reduce some of the manual production burden, content designers now have more capacity to practise those skills at a deeper level.

2. Support real skill evolution

A new role profile is not enough. People need opportunities to develop the skills, confidence, and judgement to work differently.

At Atlassian, content designers have taken part in AI boot camps, innovation weeks, tooling explorations, and hands-on work in repositories and code editors. They’ve experimented with system prompting, content modelling, AI tool development, evaluation, and visual generation.

One example came from our Senior Content Designer, Hannah Sanderson, who worked on Remix — a feature in Confluence that lets people turn text into charts, infographics, and other visual outputs.

The challenge was unfamiliar territory for content design. How do you write your way into a consistent, high-quality visual output when there is no front-end component library to control the result? Instead, a non-deterministic language model interprets a system prompt and generates the visual.

Hannah explored how content design could create consistency in that environment. She developed guidance for:

• The visual “DNA” of the prompt
• Compositional rules
• Surface treatments
• Parameters for visual outputs
• Negative guidance describing what the model should avoid

Once she had developed and tested the approach, she shared it with other content designers. That meant others could build on sample prompts and evaluation rubrics instead of starting from scratch.

Learning together has been one of the most effective ways we’ve found to reduce anxiety around the adoption curve.

Three lessons stand out:

Demoing is not learning

Watching someone else use a tool can create inspiration, but it does not build capability. People need hands-on opportunities to try, fail, and develop their own approach.

Move at the pace of the slowest student

Fast adopters are valuable, but they cannot be the only people shaping the future. If learning is designed only for the most technically confident people, most of the craft will be left behind.

Build community into the learning model

The biggest gains come when people teach one another. A prompt, rubric, prototype, or failure becomes much more valuable when it is shared and made reusable.

3. Build trust through focused proofs of concept

Before the shift, much of our team’s capacity was absorbed by necessary BAU production work: error messages, modals, labels, and other UX flows required before a feature could ship. That work was important, but it typically left little room to develop new skills or work on higher-value systems.

We began asking:

How could AI give content designers time back to work more strategically?

In 2025, a small team of three started investigating which repeatable content tasks might be suitable for AI transformation. Rather than attempting to change everything at once, they chose discrete, timeboxed pilots with measurable outcomes.

The first was a release-notes agent.

At Atlassian, release notes are often created under significant time pressure, and completing them can be a dependency for a feature release. Content designers were spending around 5% of their time supporting this work, and the function could sometimes be perceived as a blocker.

Release notes were a good candidate for an initial pilot because the content type was relatively well-defined. The team built an agent to draft them, then measured both time on task and quality with human review.

The result was strong enough to scale. A live production pipeline run by a new Content Review Desk staffed by content designers now supports around 150 release notes a month, while reducing time on task by 88%.

Other tools, including our bespoke Content Assistant Agent that has since been used thousands of times by product managers, designers, and engineers. The assistant has now been invoked more than 14,000 times.

These results did more than save time. They changed how the business understood where content design fits in the new era. Each successful experiment helped show the business that content designers could lead meaningful AI transformation that had a bottom line impact.

Some lessons we learned:

• Start with a narrow, real problem.
• Decide what you will measure before you begin.
• Use evidence to improve the solution and share wins frequently.
• Connect the work to outcomes business leaders already care about.
• Scale only when the results justify it.

The early proofs of concept have now led to broader work on the Atlassian Content System: infrastructure for producing dynamic, consistent, and trustworthy Atlassian-branded content at scale.

4. Create the psychological safety to change

The final pillar is the most difficult: helping people change in an environment where not everything can be controlled.

Much of the threat of AI comes from ambiguity about what it means for our work and whether our skills remain valuable.

As leaders, our role has been to help content designers feel like agents of their own future. That has meant:

• Setting more targeted goals and building our muscle to be business-impact focussed
• Creating opportunities to stretch into new areas
• Rethinking team structures and engagement models
• Sharing what we know is changing — and what remains unknown
• Gathering regularly as a community to celebrate and learn from each other
• Making space for people to challenge and co-design the direction

We’ve used Town Halls to give the wider craft a shared view of what is shipping, what is changing, and what we still don’t know. Team offsites have helped rebuild trust and create space for deeper collaboration.

Community consultations have also been important. These sessions, led by individual contributors, have given people a chance to ask difficult questions, challenge plans, and propose what should happen next.

We’ve tried to be direct about where we are through regular surveys and action plans. The work is not instant, but it gives people a way to see that their feedback is being heard and acted on.

One content designer recently described loving the prompt work they were doing while still missing some of the more freeform, creative moments of the craft.

That tension is real. The goal is not to replace the parts of content design that people love with technical work. It is to expand the range of work content designers can do — and help people find a sustainable balance between craft, systems, creativity, and technology.

Signals to look out for

If you’re leading content design through AI disruption, here are some things to keep in mind.

Identity

Are people defending old ways of working, or are they helping co-design what comes next? How can you support them to shift to the latter?

Skills

Do people have hands-on opportunities to build capability, or are they only watching demos and hearing strategy? What would change this?

Evidence

Are you running small experiments that demonstrate value in terms your organisation understands? (TIP: Make sure everyone hears about it!)

Space to grow

Clear communication, exemplary leadership, and equitable recognition lay the foundation for navigating organizational change effectively:

• Foster transparent communication: Ensure team members understand what is changing, what remains uncertain, and how they can actively shape the direction forward.
• Model the mindset: Leadership behavior across all tiers—executives, managers, and senior individual contributors—directly drives the emotional climate during transitions, as mindsets spread quickly.
• Recognize quiet contributors: Identify and fairly reward proactive talent. Innovation often originates away from the spotlight through individuals quietly experimenting, supporting colleagues, or pioneering emerging practices.

It’s time to write what comes next!

We don’t have a perfect playbook and we are still experimenting, learning, and changing course.

But we have seen that a vulnerable craft can become a strategic powerhouse when leaders combine:

• A stronger identity
• Practical opportunities to develop new skills
• Evidence from focused experiments
• The psychological safety to try, question, and grow

Whether you adopt all four pillars or begin with just one, intentional action can help content designers move from reacting to AI disruption to shaping the systems around it.

Content designers are not being replaced by AI. We are becoming the people who help architect how AI communicates, behaves, and creates value.

Even in the line of fire, content design has moved past just surviving and is learning how to thrive.