For the entire history of computing, enterprise applications have forced us into functional silos to get things done. Book a vacation? Log into your HCM app. Create a document? Go to your word processor. Write some code? Fire up your code editor. Today, that's no longer true, thanks to a combination of increasingly sophisticated browser technologies and the leveling effect of AI. There was a powerful demonstration of this trend during the keynote at last week's annual Config conference, when digital design vendor Figma unveiled the ability to work with code from within its flagship design canvas. Loredana Crisan, Figma's Chief Design Officer, puts this capability into context:

We're calling this Code Layers, and for the first time, basically, design and code can live in one space, side-by-side. You can generate code from a design, you can prompt for it directly with agents, and you can import it from a repository in your GitHub.

Once it's on the canvas, teams can explore and build with it, just like they would with design. This is how code becomes a creative material. With this approach, design and code are no longer a trade-off. You can move fluidly between them.

There were plenty more breakthroughs announced at the event — adding complex shading effects, perspective depth, full-motion animation and generative workflows to the canvas — each significant in themselves from a design point of view. But it was the obliteration of the historical divide between design and code that truly brought home the extent to which today's technology is blurring the barriers that have traditionally separated enterprise processes. Figma's customers design many of today's familiar applications, websites and digital tools, from Airbnb, Netflix and Uber, to Accenture, Google and Microsoft. Bringing together the separate activities of design and coding into a single canvas cuts out all the back-and-forth and miscommunication that occurs when developers and designers work in separate silos, while enabling much more experimentation and testing at the ideation stage.

Figma takes this even further by baking team collaboration into its platform. This is not just about designers being able to work directly with code instead of handing off tasks to developers and waiting for the results to come back. By embedding the ability to import, run and edit code in a shared design canvas, Figma enables real-time discussion between designers and developers, so that they can build with the benefit of each other's expertise and intuition. In the opening keynote, Nikolas Klein, Product Manager for Figma Make, the vendor's AI coding tool, explains the impact:

While agents have made it so much easier to write thousands of lines of code, the workflows around code are not built for exploring together at all. We're working in silos. Each one of us is hundreds of prompts deep with our own agents. And then at some point we look up and we throw prototype recordings at each other, hoping that they'll inspire us, only to realize that we're not working together anymore...

Designing with vectors and graphics has moved away from being done in isolation and became this beautiful collaborative activity. I believe that we will be experiencing the same shift with code.

It's worth bearing this perspective of collaboration in mind when considering the other new features Figma announced last week:

• Shaders — a new set of reusable generative plug-ins bring a range of shader effects and fills, which users can customize through a conversational interface, or by adjusting the underlying controls.
• Weave — the integration of pre-built node-based workflows from Figma Weave, providing repeatable frameworks for making and editing imagery, video, audio, and 3D effects.
• Motion — a new animation timeline, based on last year's acquisition of Modyfi, provides control over a variety of animation styles applied to any element of a design.

Figma's keynote demos made all of this look easy, but there's a huge amount of engineering going on under the surface to, for example, allow several designers to simultaneously view an animation running in the same canvas, or have several separate code bases running within a shared design while developers and designers exchange comments about their pros and cons. Marcel Weekes, VP of Product Engineering at Figma, tells me:

The user experience is tied very closely to speed. It has to feel as if that person is literally right next to you, using the same screen, and you're almost looking at the same thing. But it is really a challenge, and we spend a lot of time and effort working through that.

We have to keep it very simple, actually, in order to leverage the same multiplayer primitives that we have for Figma Design that we bring across to Slides, so that we can leverage it again on Make, as an example. The interfaces are very simple. It's complex to make it look simple.

The multiplayer capability also applies to Figma's AI agent, now generally available, and the reusable skills and context that form part of its prompts. Users can share these assets and prompts with teammates, and their conversations with agents are visible to the team by default, so that colleagues can learn from seeing what others are doing. Crisan explains why this matters:

Customers tell us that AI made individual work really easy, but collective work absolutely impossible. As a result, their teams get stuck in reviews and struggle to stay on the same page as they build...

We're adding the ability for designers to browse and build on each other's agent work. Maybe someone found a better way to synthesize feedback, maybe they built a workflow, or they figured out a new layout tool. Their discovery becomes your starting point, and this is the part of collaborating in Figma. Now, over time, some of these workflows become just how your team operates. This is why we're introducing skills. Skills let you encode your team's expertise directly into the agent — essentially, you teach Figma how your team works, and this is how everyone gets smarter together.

In its own use of this feature, Figma is finding that its designers value this ability to have the agent to create a repeatable process from work that they do every day and then share the results with colleagues. She adds:

We've been dogfooding this in our team for a couple of months, and we've realized that now we don't necessarily just think about, 'What can I create?, but, 'What tool can I create that would give me the output that I want?' So, taken together, this creates a canvas that gets more intelligent the more you use it.

I fear many vendors are making the mistake of viewing AI as a personal productivity tool rather than seeing the potential for it to lift the capabilities of teams and organizations as a whole. While it's harder to engineer agents and prompts that can be shared — especially while respecting the varied permission levels of participants — there's so much value in allowing people to learn from each other and build on their communal efforts that it's worth the extra effort.

With teamwork being such a fundamental part of its DNA, there's no danger of Figma making this mistake. On the contrary, the vendor continues to stretch the envelope of what's achievable in a shared canvas, helping designers and developers build faster and better together. Over the next few days I'll be digging into the experiences of customers and Figma's approach in further stories from last week's event — watch this space.