AI is advancing fast — and most of us are learning as we go.

For many, AI in the workplace still feels like a black box: exciting, a little mysterious, and constantly evolving. There’s no universal playbook, and everyone is experimenting in their own way.

At Observable, we believe the human side of adopting AI workflows matters most. Technology alone doesn’t create impact — people do. This is why we favor a human-in-the-loop approach to AI in Observable Canvases and why, as the People Ops leader at Observable, I’m focused on creating a company culture where curiosity thrives, experimentation feels safe, and AI naturally integrates into our workflows instead of disrupting it.

This approach isn’t about chasing the latest tool or mandating its use. It’s about giving people the space to explore, learn, and grow together. Here’s how we’re making that happen as a company:

Starting with curiosity, not control to encourage AI adoption

As a data company, we decided that we needed more data points to inform a potential AI strategy. So, we started with a question: What is the role of AI in people’s workflows at Observable today?

To find out, we launched a survey asking employees which AI tools they were using, how often, and for what purpose. When we analyzed the data,the answers showed that experimentation was already happening.

We learned that some people were using AI daily; others experimented as needed. The key insight? AI adoption was already happening organically. Our job isn’t to push it but to support it. We also gained deeper insights into which tools teams are using, including ChatGPT, Claude, Google Gemini, GitHub Copilot, Zed, Cursor, Granola, Figma AI, and NotebookLM to:

  • Draft and revise messaging

  • Generate spreadsheet functions and debug code

  • Summarize meeting notes and synthesize customer feedback

  • Scaffold new features and even create podcast-style summaries for deeper engagement

This survey is now part of our ongoing learning process. We’ll continue running it quarterly so we can track what’s changing, spot new opportunities, and keep adapting as both our needs and the technology evolve.

Removing barriers with the AI stipend

From the survey results, we also realized that people were eager to experiment but faced small hurdles, such as limited access to tools or uncertainty about which ones to try.

To remove that friction, we introduced a $50 per month AI stipend, available with manager approval. It’s intentionally simple: employees can choose tools that make sense for their work without worrying about budget constraints.

This stipend had a positive side-effect as it sparked organic sharing of recommendations on AI workflows and automation. Team members started threads comparing notes on what was worth trying, which tools were approved for broader use, and what made their work easier. These conversations have been invaluable for identifying high-impact tools and building a company culture of peer learning.

The AI stipend is a relatively small investment for our company, but the payoff has been big:

  • People feel empowered to explore AI tools without hesitation

  • Teams are learning from one another and sharing real results

  • Experimentation is translating into tangible skill-building and confidence

Embedding AI into everyday workflows

Exploration is important, but AI’s real value lies in its natural integration into day-to-day work. When getting started with AI, it can be intimidating to know where to start. At Observable, we’ve observed that even minor changes can have a significant impact on our workflows. For instance, we began using AI to summarize meeting notes. It’s a small shift, but it’s reduced the burden of manual follow-ups, clarified next steps, and helped cross-functional teams stay aligned.

These lightweight integrations show what’s possible: AI isn’t replacing human connection — it’s supporting it by automating a repetitive, administrative burden. This allows everyone in the meeting to be fully engaged and gives us back time.

We’re intentionally taking our time before introducing policies or standardized tools. For now, we’re focused on observing what’s working in practice, learning from it, and using those insights to shape future guidance.

Keeping AI adoption human-centered

With all the buzz around AI, it’s tempting to focus only on speed and efficiency. But at Observable, we’re committed to keeping people at the center.

That means giving employees room to figure things out, welcoming feedback and even “I’m still learning” moments, and encouraging open sharing of both successes and missteps. We know adoption will happen at different speeds, and that’s perfectly okay.

We’ve also been clear: AI isn’t mandatory. The tools aren’t perfect, and not every use case is obvious right away. What matters most is that people feel safe to explore on their own terms.

What’s next: Continuing to embrace AI tools

We’re continuing to learn alongside our team, and our next steps include:

  • Running our quarterly AI survey to track adoption and sentiment

  • Hosting “lunch and learns” to share practical use cases

  • Identifying more areas where AI can reduce friction or free up time

  • Developing a flexible set of AI principles

  • Encouraging employees to attend AI events and bringing back what they learn

Our vision is for AI to be a natural part of how we work, not something separate or intimidating.

Final thoughts: Cultivating conditions to enable organic AI adoption

Building an AI-driven company culture isn’t about adopting the shiniest tool or rushing to keep up with trends. It’s about creating the conditions where people feel empowered to explore, share, and learn together.

HR plays an essential role in making that possible. By lowering barriers, fostering psychological safety, and keeping our focus on people, we are working to ensure that AI is not just powerful but truly human-centered.

We’re still learning, and we’d love to hear from you: How is your organization approaching AI adoption? What’s working for your teams?