Design leaders must help teams adopt AI thoughtfully, not reactively.
Without guidance, teams adopt AI randomly. Everyone uses ChatGPT differently. Some refuse to touch it. Others over-rely on it.
In the beginning it’s normal that everyone experiments independently. There’ll be no shared understanding of what works, no evaluation criteria, and no quality standards.
Where you want to get is to have clear uses cases, evaluation criteria and, most importantly, a shared understanding of how to use AI and why. Teams need to learn together what works and what doesn’t
Left alone, most teams never move to this second phase.
Define use cases. Where does AI help? Research synthesis. First-draft documentation. Exploring alternatives. Generating variants. Be specific.
Set quality standards. AI output is input to your process, not the final product. Define what “good enough from AI” means. What requires human refinement?
Establish evaluation criteria. How do you judge AI outputs? What makes a good prompt? When is AI the right tool? When is it a wrong tool?
Share learnings. When someone discovers effective AI use, share it. When AI fails, discuss why. Build collective knowledge.
Address fears directly. Some designers fear AI replaces them. Address this explicitly. Use AI to augment human autonomy, not replace it. Be clear about this.
You can’t ban AI, you can’t ignore it
You can’t mandate how people use AI. You can’t ban it (they’ll use it anyway). You can’t ignore it.
You must guide adoption through:
- Clear communication about use cases
- Demonstration of effective practices
- Discussion of failures and limitations
- Setting quality expectations
- Creating space for experimentation
This connects to Good leaders lay the breadcrumbs. You don’t micromanage AI usage. You show the path, set standards, and let teams find what works.
For remote teams
Remote first teams need this more. You can’t observe what people are doing with AI. You can’t have hallway conversations about what works.
Async communication becomes critical. Document AI use cases. Share prompts that work. Discuss failures publicly. Build shared knowledge through writing.
Documentation as a remote leadership tool applies here. Your written guidance on AI adoption becomes the reference point. Clear documentation replaces synchronous training. Documentation matters more in an AI world.
The job of a manager isn't a creative one. Your job is creating conditions for good work. Thoughtful AI adoption is part of that.