Financial infrastructure platform | 8,000+ employees
Financial infrastructure platform | 8,000+ employees
Financial infrastructure platform | 8,000+ employees
summary
How Stripe designs with AI
What happens when you create the conditions for AI adoption instead of prescribing a playbook?
Stripe has long operated on a simple principle: hire people with high agency, then trust them to figure out how they work best. When AI tooling arrived, they didn't replace that philosophy with mandates or standardized workflows. Instead, they made experimentation easier, letting designers shape how AI fit into their own workflows.
What happened went beyond adoption. Designers didn't just learn AI tools; they built their own. By turning recurring workflow pain points into custom internal tools, the team was able to embed its own standards and context directly into the way work gets done. And as those tools spread across the organization, even beyond the design team, they helped raise the quality bar for everyone.

summary
How Stripe designs with AI
What happens when you create the conditions for AI adoption instead of prescribing a playbook?
Stripe has long operated on a simple principle: hire people with high agency, then trust them to figure out how they work best. When AI tooling arrived, they didn't replace that philosophy with mandates or standardized workflows. Instead, they made experimentation easier, letting designers shape how AI fit into their own workflows.
What happened went beyond adoption. Designers didn't just learn AI tools; they built their own. By turning recurring workflow pain points into custom internal tools, the team was able to embed its own standards and context directly into the way work gets done. And as those tools spread across the organization, even beyond the design team, they helped raise the quality bar for everyone.

summary
How Stripe designs with AI
What happens when you create the conditions for AI adoption instead of prescribing a playbook?
Stripe has long operated on a simple principle: hire people with high agency, then trust them to figure out how they work best. When AI tooling arrived, they didn't replace that philosophy with mandates or standardized workflows. Instead, they made experimentation easier, letting designers shape how AI fit into their own workflows.
What happened went beyond adoption. Designers didn't just learn AI tools; they built their own. By turning recurring workflow pain points into custom internal tools, the team was able to embed its own standards and context directly into the way work gets done. And as those tools spread across the organization, even beyond the design team, they helped raise the quality bar for everyone.

We set the table to enable the party. We’re comfortable with progress over perfection. And when rolling out a new internal AI tool, we admit it will evolve and avoid mandating it. Our goal is simply to get better over time by learning from our teams' experiences.

Katie Dill
Head of Design, Stripe
We set the table to enable the party. We’re comfortable with progress over perfection. And when rolling out a new internal AI tool, we admit it will evolve and avoid mandating it. Our goal is simply to get better over time by learning from our teams' experiences.

Katie Dill
Head of Design, Stripe
We set the table to enable the party. We’re comfortable with progress over perfection. And when rolling out a new internal AI tool, we admit it will evolve and avoid mandating it. Our goal is simply to get better over time by learning from our teams' experiences.

Katie Dill
Head of Design, Stripe
KEY INSIGHTS
What we learned
01
Give people time and permission, not a mandate
Rather than standardizing AI adoption around a single tool or workflow, Stripe focused on removing the obstacles to experimenting. Designers get dedicated time through AI-cation days, dozens of approved tools to choose from, Fireside sessions where people demo what they've built, and a leaderboard highlighting popular internal plugins. As Katie put it, "it's more through carrots than sticks." The goal isn't to tell people how to use AI. It's to make trying new things easier.
02
Build tools that encode your standards
As AI lowers the cost of building software, more people can create internal tools. Stripe's philosophy is to let those tools emerge from the people closest to the problem. ProtoDash wasn't commissioned; it grew out of one designer's frustration with repetitive prototyping. Dante came from the content team trying to maintain quality at scale. Together, they illustrate a broader shift within the company: instead of simply documenting best practices, they increasingly build them into the tools themselves.
03
Leaders should walk the talk
Katie's advice to design leaders is simple: "walk the talk." Build something with AI. Ship a side project. Better yet, use the tools in work that ships to customers. Visible participation signals that experimentation isn't extracurricular; it's a core part of your job as a designer. It also leads to better decisions, as leaders who've experienced the strengths and frustrations firsthand are much better equipped to decide which tools deserve broader adoption.
04
Make small bets before you make big ones
Stripe calls this "roof shots, not moonshots." Rather than betting big on one approach up front, the team tries a lot of small, low-effort things, sees what people actually keep using, and only makes a company-wide bet (training everyone on one tool, standardizing on one approach) once something's proven itself. "I think that's a good approach for a company of any size," Katie says, since betting early on one tool or approach is risky when the tools themselves keep changing so quickly.
KEY INSIGHTS
What we learned
01
Give people time and permission, not a mandate
Rather than standardizing AI adoption around a single tool or workflow, Stripe focused on removing the obstacles to experimenting. Designers get dedicated time through AI-cation days, dozens of approved tools to choose from, Fireside sessions where people demo what they've built, and a leaderboard highlighting popular internal plugins. As Katie put it, "it's more through carrots than sticks." The goal isn't to tell people how to use AI. It's to make trying new things easier.
02
Build tools that encode your standards
As AI lowers the cost of building software, more people can create internal tools. Stripe's philosophy is to let those tools emerge from the people closest to the problem. ProtoDash wasn't commissioned; it grew out of one designer's frustration with repetitive prototyping. Dante came from the content team trying to maintain quality at scale. Together, they illustrate a broader shift within the company: instead of simply documenting best practices, they increasingly build them into the tools themselves.
03
Leaders should walk the talk
Katie's advice to design leaders is simple: "walk the talk." Build something with AI. Ship a side project. Better yet, use the tools in work that ships to customers. Visible participation signals that experimentation isn't extracurricular; it's a core part of your job as a designer. It also leads to better decisions, as leaders who've experienced the strengths and frustrations firsthand are much better equipped to decide which tools deserve broader adoption.
04
Make small bets before you make big ones
Stripe calls this "roof shots, not moonshots." Rather than betting big on one approach up front, the team tries a lot of small, low-effort things, sees what people actually keep using, and only makes a company-wide bet (training everyone on one tool, standardizing on one approach) once something's proven itself. "I think that's a good approach for a company of any size," Katie says, since betting early on one tool or approach is risky when the tools themselves keep changing so quickly.
KEY INSIGHTS
What we learned
01
Give people time and permission, not a mandate
Rather than standardizing AI adoption around a single tool or workflow, Stripe focused on removing the obstacles to experimenting. Designers get dedicated time through AI-cation days, dozens of approved tools to choose from, Fireside sessions where people demo what they've built, and a leaderboard highlighting popular internal plugins. As Katie put it, "it's more through carrots than sticks." The goal isn't to tell people how to use AI. It's to make trying new things easier.
02
Build tools that encode your standards
As AI lowers the cost of building software, more people can create internal tools. Stripe's philosophy is to let those tools emerge from the people closest to the problem. ProtoDash wasn't commissioned; it grew out of one designer's frustration with repetitive prototyping. Dante came from the content team trying to maintain quality at scale. Together, they illustrate a broader shift within the company: instead of simply documenting best practices, they increasingly build them into the tools themselves.
03
Leaders should walk the talk
Katie's advice to design leaders is simple: "walk the talk." Build something with AI. Ship a side project. Better yet, use the tools in work that ships to customers. Visible participation signals that experimentation isn't extracurricular; it's a core part of your job as a designer. It also leads to better decisions, as leaders who've experienced the strengths and frustrations firsthand are much better equipped to decide which tools deserve broader adoption.
04
Make small bets before you make big ones
Stripe calls this "roof shots, not moonshots." Rather than betting big on one approach up front, the team tries a lot of small, low-effort things, sees what people actually keep using, and only makes a company-wide bet (training everyone on one tool, standardizing on one approach) once something's proven itself. "I think that's a good approach for a company of any size," Katie says, since betting early on one tool or approach is risky when the tools themselves keep changing so quickly.



in action
ProtoDash: Turning prompts into production-ready prototypes
RYAN SPENCER AND SADHIKA BILLA, STAFF PRODUCT DESIGNERS
Owen Williams kept running into the same problem: AI-generated prototypes looked generic, with the wrong fonts, components, and layouts. So he built ProtoDash as a side project—a browser-based prototyping studio that turns prompts into working React prototypes built with Sail, Stripe's internal design system. It quickly spread across the design team, becoming one of Stripe's most widely used internal AI tools.
What's cool about this is that it was built as a passion project from one of our product design managers, Owen. It wasn't a mandated thing. He just decided to build this to make it a lot easier for not just designers, but also engineers and product managers.

Sadhika Billa
STAFF PRODUCT DESIGNER, STRIPE
Ryan Spencer and Sadhika Billa used ProtoDash to generate a live fraud dashboard from a single prompt. Instead of a static mockup, they got a functional prototype complete with fraud rate, blocked payments, protected revenue, and risk-level breakdowns—all built with Stripe's production UI components and ready to iterate on.
in action
ProtoDash: Turning prompts into production-ready prototypes
RYAN SPENCER AND SADHIKA BILLA, STAFF PRODUCT DESIGNERS
Owen Williams kept running into the same problem: AI-generated prototypes looked generic, with the wrong fonts, components, and layouts. So he built ProtoDash as a side project—a browser-based prototyping studio that turns prompts into working React prototypes built with Sail, Stripe's internal design system. It quickly spread across the design team, becoming one of Stripe's most widely used internal AI tools.
What's cool about this is that it was built as a passion project from one of our product design managers, Owen. It wasn't a mandated thing. He just decided to build this to make it a lot easier for not just designers, but also engineers and product managers.

Sadhika Billa
STAFF PRODUCT DESIGNER, STRIPE
Ryan Spencer and Sadhika Billa used ProtoDash to generate a live fraud dashboard from a single prompt. Instead of a static mockup, they got a functional prototype complete with fraud rate, blocked payments, protected revenue, and risk-level breakdowns—all built with Stripe's production UI components and ready to iterate on.
in action
ProtoDash: Turning prompts into production-ready prototypes
RYAN SPENCER AND SADHIKA BILLA, STAFF PRODUCT DESIGNERS
Owen Williams kept running into the same problem: AI-generated prototypes looked generic, with the wrong fonts, components, and layouts. So he built ProtoDash as a side project—a browser-based prototyping studio that turns prompts into working React prototypes built with Sail, Stripe's internal design system. It quickly spread across the design team, becoming one of Stripe's most widely used internal AI tools.
What's cool about this is that it was built as a passion project from one of our product design managers, Owen. It wasn't a mandated thing. He just decided to build this to make it a lot easier for not just designers, but also engineers and product managers.

Sadhika Billa
STAFF PRODUCT DESIGNER, STRIPE
Ryan Spencer and Sadhika Billa used ProtoDash to generate a live fraud dashboard from a single prompt. Instead of a static mockup, they got a functional prototype complete with fraud rate, blocked payments, protected revenue, and risk-level breakdowns—all built with Stripe's production UI components and ready to iterate on.



in action
Dante: Embedding Stripe's writing standards into AI
CHRIS GREER, STAFF CONTENT DESIGNER
Chris Greer on the content design team built Dante with an engineering PM after realizing the content team could no longer manually review everything Stripe shipped. Rather than relying on a final editorial pass, Dante checks copy against Stripe's style guide throughout development—from Slack to GitHub to the command line—helping teams catch issues long before launch. Instead of replacing editorial judgment, it scales it, ensuring Stripe's voice stays consistent as the company moves faster.
We don't want to let the quality bar slip just because we're shipping more and faster. We still care about sentence case and en dashes and Oxford commas. We still wanna have that Stripe quality. It just became a question of designing a system to enable that at this newer scale.

Chris Greer
STAFF CONTENT DESIGNER
in action
Dante: Embedding Stripe's writing standards into AI
CHRIS GREER, STAFF CONTENT DESIGNER
Chris Greer on the content design team built Dante with an engineering PM after realizing the content team could no longer manually review everything Stripe shipped. Rather than relying on a final editorial pass, Dante checks copy against Stripe's style guide throughout development—from Slack to GitHub to the command line—helping teams catch issues long before launch. Instead of replacing editorial judgment, it scales it, ensuring Stripe's voice stays consistent as the company moves faster.
We don't want to let the quality bar slip just because we're shipping more and faster. We still care about sentence case and en dashes and Oxford commas. We still wanna have that Stripe quality. It just became a question of designing a system to enable that at this newer scale.

Chris Greer
STAFF CONTENT DESIGNER
in action
Dante: Embedding Stripe's writing standards into AI
CHRIS GREER, STAFF CONTENT DESIGNER
Chris Greer on the content design team built Dante with an engineering PM after realizing the content team could no longer manually review everything Stripe shipped. Rather than relying on a final editorial pass, Dante checks copy against Stripe's style guide throughout development—from Slack to GitHub to the command line—helping teams catch issues long before launch. Instead of replacing editorial judgment, it scales it, ensuring Stripe's voice stays consistent as the company moves faster.
We don't want to let the quality bar slip just because we're shipping more and faster. We still care about sentence case and en dashes and Oxford commas. We still wanna have that Stripe quality. It just became a question of designing a system to enable that at this newer scale.

Chris Greer
STAFF CONTENT DESIGNER
in action
Stripe Press: Getting to the good ideas faster
PABLO DELCAN, DESIGNER, STRIPE PRESS
Before AI, Pablo Delcan sketched book covers by hand, assembling references from stock libraries, found imagery, and scanned books. A single concept could take 20 minutes or more just to determine whether it was worth pursuing. He was initially hesitant to bring AI into that process, worried it would replace the part of the work he loved most. Instead, it removed the slowest part. Today he explores dozens of directions in parallel—testing different materials, foils, and embossing before committing to physical samples, a process that once took weeks. The concept is still his. The final cover is still his. AI simply compresses the stretch between the two.
The concept still comes from us. The final cover is still made by us. AI just compressed the middle part of the process, the part where I used to spend days finding out whether an idea was worth the effort.

Pablo Delcan
DESIGNER, STRIPE PRESS
in action
Stripe Press: Getting to the good ideas faster
PABLO DELCAN, DESIGNER, STRIPE PRESS
Before AI, Pablo Delcan sketched book covers by hand, assembling references from stock libraries, found imagery, and scanned books. A single concept could take 20 minutes or more just to determine whether it was worth pursuing. He was initially hesitant to bring AI into that process, worried it would replace the part of the work he loved most. Instead, it removed the slowest part. Today he explores dozens of directions in parallel—testing different materials, foils, and embossing before committing to physical samples, a process that once took weeks. The concept is still his. The final cover is still his. AI simply compresses the stretch between the two.
The concept still comes from us. The final cover is still made by us. AI just compressed the middle part of the process, the part where I used to spend days finding out whether an idea was worth the effort.

Pablo Delcan
DESIGNER, STRIPE PRESS
in action
Stripe Press: Getting to the good ideas faster
PABLO DELCAN, DESIGNER, STRIPE PRESS
Before AI, Pablo Delcan sketched book covers by hand, assembling references from stock libraries, found imagery, and scanned books. A single concept could take 20 minutes or more just to determine whether it was worth pursuing. He was initially hesitant to bring AI into that process, worried it would replace the part of the work he loved most. Instead, it removed the slowest part. Today he explores dozens of directions in parallel—testing different materials, foils, and embossing before committing to physical samples, a process that once took weeks. The concept is still his. The final cover is still his. AI simply compresses the stretch between the two.
The concept still comes from us. The final cover is still made by us. AI just compressed the middle part of the process, the part where I used to spend days finding out whether an idea was worth the effort.

Pablo Delcan
DESIGNER, STRIPE PRESS



FEATURED STRIPE TEAM MEMBERS
NAME
POSITION
Katie Dill
Head of Design
Ryan Spencer
Staff Product Designer
Sadhika Billa
Staff Product Designer
Chris Greer
Staff Content Designer
Pablo Delcan
Designer, Stripe Press
FEATURED STRIPE TEAM MEMBERS
NAME
POSITION
Katie Dill
Head of Design
Ryan Spencer
Staff Product Designer
Sadhika Billa
Staff Product Designer
Chris Greer
Staff Content Designer
Pablo Delcan
Designer, Stripe Press
FEATURED STRIPE TEAM MEMBERS
NAME
POSITION
Katie Dill
Head of Design
Ryan Spencer
Staff Product Designer
Sadhika Billa
Staff Product Designer
Chris Greer
Staff Content Designer
Pablo Delcan
Designer, Stripe Press
Join the team
Stripe Design
Stripe's design organization brings together a multidisciplinary team across product, brand, content, and motion design, research, and engineering—all working toward a shared mission: making economic infrastructure simple and accessible. The team combines a high bar for craft with a culture of curiosity and experimentation, creating space for designers to rethink how they work as new possibilities emerge. The team is hybrid, with hubs in San Francisco, New York, Seattle, Dublin, and London.
Join the team
Stripe Design
Stripe's design organization brings together a multidisciplinary team across product, brand, content, and motion design, research, and engineering—all working toward a shared mission: making economic infrastructure simple and accessible. The team combines a high bar for craft with a culture of curiosity and experimentation, creating space for designers to rethink how they work as new possibilities emerge. The team is hybrid, with hubs in San Francisco, New York, Seattle, Dublin, and London.
Join the team
Stripe Design
Stripe's design organization brings together a multidisciplinary team across product, brand, content, and motion design, research, and engineering—all working toward a shared mission: making economic infrastructure simple and accessible. The team combines a high bar for craft with a culture of curiosity and experimentation, creating space for designers to rethink how they work as new possibilities emerge. The team is hybrid, with hubs in San Francisco, New York, Seattle, Dublin, and London.
FURTHER READING
FURTHER READING
FURTHER READING

