Skip to main content

What it takes to build great products now

Getting to a prototype is easier than ever. The harder work is sustaining speed all the way to launch, choosing the right direction, and creating products that stand out. Here’s what that takes.

Share What it takes to build great products now

Illustration by Lena Muller

With AI, you can go from writing a prompt to something clickable in a few hours. But getting there fast doesn’t mean you've built something great. A polished prototype doesn’t get you past the big questions: Is this direction the best or just the first? Does it feel like us? And above all: Is it actually worth building?

The teams answering those questions well are strong in three areas—speed, direction, and differentiation. They keep moving through the final stretch, weigh many directions before betting on one, and shape the result into something only they would make.

Speed: Keep up momentum until you ship

When the first 80% is easy, teams can lose momentum in the last 20% of the work because a prototype can feel like something the team can actually launch. Product builders might realize that the prototype they made quickly doesn’t actually use their design system or creates extra work when it moves from design into code. The challenge is holding onto that early speed

as the work gets more detailed, involves more people, and gets closer to shipping.

Colorful background with the word "Chaos as a catalyst: Building Sephora's next operating model" written on top.Colorful background with the word "Chaos as a catalyst: Building Sephora's next operating model" written on top.

Learn how Sephora SVP of Technology Sneha Narahalli uses speed for more than just shipping fast.

Use the fastest medium for the job

As products take shape, teams iterate constantly, from small visual tweaks to changes in how a feature behaves. The fastest approach depends on what you’re solving for. For example, if there’s too much space between a form field and a button, it may be quicker to adjust the spacing directly in Figma than to describe the change in a prompt. Other changes are faster to make in code. If you’re building a calculator that lets users estimate the price based on how many teammates will use the product, you don’t need to design a separate version for 10 people, 20 people, or 50 people. You can set up a simple calculation in code so the price updates automatically as the number changes.

The biggest speed gains come from choosing the fastest way to make each kind of change, rather than forcing every task through the same workflow. In Figma, that can mean moving between direct editing, prompting, and code depending on the change you need to make.

Choose the right tool for the task:

Hand the whole team the shortcut

As work moves through the team, maintaining speed means making useful shortcuts repeatable. Say a designer figures out a fast and reliable way to check a new flow against your accessibility standards. Instead of keeping that process to themselves, they can turn it into a reusable workflow for the whole team. In Figma, teams can turn those repeatable ways of working into skills or plugins that anyone can reuse.

Skill vs. plugin

A skill is a reusable set of instructions you can trigger with a / command, like checking a design against accessibility guidelines.

A plugin is a reusable tool that performs a specific task on the canvas, like generating a dashboard layout or organizing a set of images.

Make better ways of working available to the whole team in Figma:

Direction: Go broad before you choose

Exploring multiple directions used to take longer, which sometimes forced teams to narrow quickly or keep ideas too hypothetical to really understand how (or if) they’d work. Now, the same tools that speed up a first direction also make it faster to try more alternatives and see how promising ideas hold up in more situations before committing.

Colorful background with the words "Designing OpenAI's products: Rethinking design in the age of AI" written overtopColorful background with the words "Designing OpenAI's products: Rethinking design in the age of AI" written overtop

Learn how Ian Silber, Head of Product Design at OpenAI, and his team go wide on ideas before deciding what’s worth shipping.

Check the edge cases before you commit

A single prototype hides the branches, the empty states, and the awkward transitions where even a promising concept can create problems in practice. Say you’re designing a flow for booking salon appointments. Someone picks an open slot, then gets a confirmation. But the prototype doesn’t show what happens if nothing is available that week, or when the hairstylist cancels, or when the customer needs to reschedule. Those are the moments where a prototype is too limiting to illustrate the full picture, and where you find out whether the direction actually holds.

Get the full view in Figma:

Compare directions together

Ideas are easier to judge when they have company. Putting several directions next to each other sharpens the trade-offs, like which one pushes the visual identity further and which one prioritizes a clearer UI experience. In addition, a lot of AI experimentation happens in isolation, but teams need to compare, react, disagree, and build on one another’s ideas. A shared canvas gives them a place to do that while keeping the directions and the discussion together, so feedback doesn’t get lost in a separate thread.

Alongside teammate feedback, ask the Figma agent to critique a direction through a specific lens, like first-time usability or enterprise sales concerns. That gives the team another set of questions to consider before choosing a direction.

Explore and compare directions in Figma:

Differentiation: Shape products that earn attention

When AI can get more people to a polished starting point, standing out takes more deliberate choices. Differentiation comes from pushing beyond the first output and shaping the experience until it reflects a point of view that’s distinctly yours.

Colorful background with "Scaling taste, craft, and quality in the AI era" written overtopColorful background with "Scaling taste, craft, and quality in the AI era" written overtop

Hear from Teo Connor, VP of Design at Airbnb, on how to preserve quality in a world where it’s easier than ever to go from 0-1.

Start from a strong foundation

A strong design system

carries your product’s point of view. Its components and variables encode thousands of small decisions about what belongs and what doesn’t—details like which red is on brand, how rounded a card should be, how much space belongs between elements, or which button treatment signals the primary action. When your team builds those choices into the system, AI-generated work can inherit the same visual language instead of falling back on generic defaults. The qualities that make your product distinct show up consistently in everything the team creates.

Make your standards reusable in Figma:

  • Ask the Figma agent to work from your design system and flag where the work strays from it.
  • Use Code Connect with Figma MCP to show agents which components in Figma match the components your team uses in code, so they can build with the same system instead of guessing or re-writing components.

Refine the output directly

Prompting is powerful, but making one small change through another prompt can mean regenerating more than you intended. Great work comes from taking ownership of the output: keeping what works, changing what doesn’t, and pushing on the details until the result reflects your intent. Imagine you generate a custom shader and only the glow is too intense. Ask for a fix through another prompt, and you risk getting back a version where the color, blur, and pattern have shifted too. Instead, adjust the glow’s intensity manually.

Take control in Figma:

Express more

Once you can refine the output directly, you can also push it further with visual expression like custom illustration and motion. That used to mean pulling in another tool or another specialist, but those constraints are fading. Say a designer wants a menu to slide open with a specific bounce instead of just appearing, or a card to use a shimmering gradient instead of a flat fill. Now that same designer can create both without leaving the canvas, pairing with the Figma agent to bring their vision to life.

Motion on the Figma canvas lets you build and refine animation without leaving the file, and reuse it across every instance of a component.

To use it:

Start from a preset animation style, describe what you want and let the Figma agent build a first pass, or set keyframes yourself for position, scale, rotation, or opacity.

Drag layers on the timeline to adjust timing, or jump the marker to any moment to check how it looks.

Once it’s built, a developer can copy the exact result as CSS, JSON, or React straight from Dev Mode.

Push your product experience further:

Building great products hasn’t become effortless. What’s changed is where the hard work happens. A prototype can be the opening move, but teams still have to keep momentum through the last mile, choose the right direction, and shape something distinct. Together, speed, direction, and differentiation turn faster building into a product people actually want to use.

Paige Costello is the VP of Product at Figma leading the Editor group, responsible for Figma Design. Before Figma, she led Core Product and was the Head of AI at Asana. She has also held leadership roles at Intercom and Intuit. Paige is passionate about building products customers love and growing the people who bring them to life.

Create and collaborate with Figma

Get started for free