Your Team Members Use AI. But Is Your Company AI-Native?

A few years ago, as a technical leader, I needed to make my team more agile. To start improving anything, you must first understand your current state. For instance, if you want to improve an algorithm, you first need to benchmark the metrics you've defined. Then, based on those benchmarks, you can make changes and run experiments to understand which improvements are actually effective.

This is a fundamental rule for any improvement you want to make. But sometimes it's hard to understand your current state. For example, how can you measure how agile your team really is?

After doing some research, I found an agility maturity framework that described different stages, from individual contributors to the entire organization. Each stage had its own criteria, which helped you understand where you currently stood and what you needed to improve to reach the next level.

What I found interesting was that agility wasn't just about individuals or a single team. It needed to be considered across the entire organization.

Recently, while working with different teams as a part-time collaborator or consultant on AI adoption, I realized that we're facing a very similar problem.

Using AI Tools Doesn't Make Your Company AI-Native - AI adoption isn't just about individuals using AI tools. It's also about shared goals, improving team workflows, and making changes across the entire organization.

The Same Challenge with AI Adoption

I haven't developed a comprehensive framework for measuring AI adoption, but based on my experience, I've noticed that many companies don't really know where they stand. Everyone says they want to adopt AI, but in many companies, this simply means developers using AI coding assistants or a few people using tools like Lovable to build products through vibe coding.

It's important to understand your current state because there is no golden method that you can apply to every team or company and expect immediate improvements.

For example, I once spent around four hours discussing AI adoption with a company. But later, I realized that some people hadn't understood what I was trying to explain. One of them thought I was saying that we could build an entire application in just one day, which wasn't my point at all.

This made me realize that before discussing AI tools, automation, or productivity, we first need to make sure everyone understands what we're actually trying to achieve.

Four Dimensions of AI Adoption

Based on my experience, I think we can look at AI adoption from four different dimensions: shared understanding, individual adoption, team workflows, and organization-wide adoption.

Each dimension has its own challenges, and improving only one doesn't necessarily improve the whole company.

For each dimension, I've included a question that can help you evaluate your current state and identify areas that might need improvement. These questions are just examples, not a complete assessment framework.

A. Shared Understanding and Goals

The first dimension is shared understanding and shared goals. When people in a company talk about AI, how they can use it, or what they want to achieve, everyone might have a different picture in mind, like the story of the blind men and the elephant.

For example, one person might think AI is just useful for refactoring code. Another might think it's only about vibe coding, while someone else, maybe the CEO or another team, might expect to build an entire application in one day.

That's why I think everyone needs a shared understanding of what AI can actually do, what the company's goals are, and what roadmap they should follow.

Ask yourself: Can everyone in the company explain our AI adoption goals in roughly the same way

B. Individual Adoption

The second dimension is individual adoption, where everyone can use AI to improve their performance in daily work.

For example, if you're a developer, you can use AI to write code. If you're a designer, you can use AI to create prototypes faster and test different ideas. If you're a QA engineer, you can use AI to help test applications.

I think this is the low-hanging fruit of AI adoption and one of the easiest places to start.

I don't want to go into too much detail here because it would make the article longer, but you can break this dimension into multiple levels.

For example, at the beginning, you might use a standalone tool like ChatGPT, where you manually provide the context. At another level, you can integrate AI with your codebase or other resources, allowing it to gather the necessary context itself.

You can imagine a range from 100% manual work with zero automation to 100% automation with zero manual work. But the goal isn't necessarily to reach 100% automation. Depending on the project, the problem you're solving, and the risks involved, you need to find the right balance between automation and manual work. Increasing automation sometimes requires a lot of effort. You need to understand how to optimize context for AI tools, how to create domain-specific agents for your work, and how to make these tools more effective for your particular needs.

Ask yourself: Are individuals actually improving their daily work with AI, or are they just experimenting with different tools?

Individual productivity ≠ Team productivity - Improving how quickly someone completes a task doesn't necessarily improve how quickly the team delivers value.

C. Team Workflows

The third dimension is about team workflows. In different companies, I've noticed that sometimes two or three developers use AI extensively while others barely use it.

But even if everyone on the team uses AI, the processes for managing and delivering work might still follow the same traditional approach. This can limit the benefits of individual AI adoption.

That's why I think teams need to rethink and improve their workflows to better support AI-assisted work. As a team lead, CTO, or CEO, you can't just give everyone a Claude Code account and expect significant improvements without changing how the team works.

For example, in one of my teams, before adopting AI, we had enough QA engineers to test the features manually our developers produced. Our development and testing capacity were reasonably balanced.

But when developers started using AI, they could build features much faster. The QA team's capacity hadn't changed because we couldn't hire more QAs, so they couldn't test everything at the same speed. As a result, more and more features were waiting for testing, and we couldn't deploy them as quickly as we developed them.

This experience showed me that improving individual productivity doesn't necessarily improve the performance of the whole team. AI helped us develop faster, but it also exposed a bottleneck in our workflow. To benefit from that increased productivity, we needed to rethink and improve our testing process as well. To solve the problem, we introduced a new guideline and process for how we should test features, which features need QA testing, and which do not need it.

Ask yourself: Has AI improved our overall delivery process, or has it just made individual tasks faster?

Organization-Wide Adoption

The fourth dimension is organization-wide adoption. Even if one team uses AI extensively while other teams still work in the same traditional way, you might see some improvements, but that doesn't necessarily make your company AI-native.

You also need to think about how different teams work together and how AI adoption in one team affects the others.

AI adoption can also affect how you build your company's roadmap and make decisions. When AI changes how quickly teams can deliver work, you may need to reconsider your planning process and even the metrics you use to measure performance. The decisions that made sense before adopting AI might not always be the right ones afterward.

Another important part is hiring and company culture. For example, as a CTO, you probably want to hire people who are curious, willing to learn new things, and comfortable challenging their existing approaches.

AI tools and capabilities change quickly, so people need to continuously learn and adapt. Leadership also needs to support this, because changing the tools without changing the company culture might not be enough.

That's why I believe becoming AI-native isn't just about adopting tools across the company. It's also about rethinking how teams collaborate, how leaders make decisions, and how the organization learns and adapts.

Ask yourself: Are we improving how the entire organization operates, or are improvements limited to a few teams?

Conclusion

However, I should mention that AI adoption doesn't always have to happen step by step. You can evaluate your progress in each dimension and work on improving several of them in parallel.

AI adoption is not just about using more AI tools. - It's about understanding your current state, identifying bottlenecks, and improving the organization as a whole.

For example, you can discuss these dimensions with the people responsible for different parts of the organization, understand your current state, and identify what needs improvement.

Imagine you're at 100% in one dimension but close to zero in the others. You might not see much overall improvement. But if you're at 30% across all four dimensions, you might achieve better results. These numbers are just examples, but the point is that improving the entire workflow can be more valuable than optimizing only one part of it.