Deck is an AI-powered data infrastructure startup that builds browser-based agents to automatically extract user-permissioned data from websites without the need for APIs. Pulse 2.0 interviewed Deck founder and CEO Yves-Gabriel Leboeuf (YG) to learn more about the company.
YG’s Background
Could you tell me more about your background? YG said:
I have spent most of my career building technology companies focused on infrastructure, software, and the systems that enable businesses to operate more efficiently.
I started building businesses at a young age, and over time, I became increasingly interested in the underlying technology that allows companies to scale. My work has consistently centered around solving problems related to access, connectivity, and automation, especially in areas where existing systems create friction for businesses and their customers.
A recurring theme throughout my career has been seeing how much innovation can happen when previously difficult-to-access systems become more programmable. When new infrastructure becomes available, it often creates opportunities for entirely new categories of products and companies.
That perspective shaped how I viewed the rise of AI. As AI models became increasingly capable, it became clear that the next challenge was not only improving intelligence, but enabling those systems to interact with the software environments where businesses actually operate.
Today, I’m focused on building the infrastructure that allows AI agents to securely access, navigate, and complete workflows across the complex software systems businesses rely on every day.
Formation Of The Company
How did the idea for the company come together? YG shared:
The idea for Deck came from a simple observation. AI was advancing much faster than the software ecosystem around it.
AI models became incredibly capable at understanding information, reasoning through problems, and generating outputs. However, most business software was still designed around human interaction. Important workflows remained locked behind logins, permissions, portals, and interfaces that were never built for autonomous systems.
I started thinking about what would be required for AI agents to move beyond providing answers and actually complete work. The missing piece was infrastructure that could connect these intelligent systems with the existing software landscape.
That became the foundation for Deck. My co-founder and I built an execution layer that allows AI agents to securely authenticate, navigate applications, and complete workflows across systems that may not have APIs or modern integrations.
My role has focused primarily on the broader company vision, strategy, and helping define where Deck fits within the evolution of AI infrastructure. I spend a significant amount of time working with customers, understanding the challenges they face deploying AI into production, and ensuring we are building the right foundation for the future. The opportunity we see is that AI adoption will not happen by replacing every existing system overnight. It will happen by making the software businesses already depend on far more accessible and useful.
Favorite Memory
What has been your favorite memory working for the company so far? YG reflected:
One of my favorite moments building Deck is a recurring one, and that is seeing customers take the technology and solve the exact problems that they come to us with.
When we started the company, we believed there was a major gap between what AI models could understand and what businesses actually needed them to do. Seeing that idea translate into results has been incredibly rewarding.
For example, Rampart needed a way to access invoices and purchase orders stored across hundreds of vendor portals. The challenge wasn’t just missing APIs. Many of those portals actively blocked automation through anti-bot protections and required 2FA at every login, meaning a human had to stay in the loop just to approve authentication. With Deck, Rampart was able to fully automate those workflows, with one developer shipping around 20 integrations in roughly half a day while avoiding significant engineering effort.
What has been most exciting is seeing customers discover new use cases once they have reliable access to these systems. That is when you realize you are not just solving one workflow problem, you are building infrastructure that enables companies to rethink what is possible with AI.
Core Products
What are the company’s core products and features? YG explained:
At its core, Deck is infrastructure that enables AI agents to securely interact with the software systems businesses rely on every day.
A lot of the AI conversation has focused on intelligence. How well models can reason, generate content, or understand information. The next challenge is enabling those systems to reliably take action.
Most businesses operate across a complex mix of applications, portals, and legacy systems. Many of these systems contain valuable data and workflows but were designed exclusively for human users. In fact, approximately 95% of our deployment projects involve software without APIs.
Deck provides the execution layer between AI agents and those systems. Our platform allows agents to securely authenticate, manage sessions, navigate interfaces, retrieve information, submit forms, update records, and complete multi-step workflows. The goal is to make the existing software ecosystem programmable for AI.
Instead of requiring companies to rebuild their technology infrastructure, Deck allows them to unlock the systems they already use and transform manual processes into reliable AI-powered workflows.
Challenges Faced
Have you faced any challenges in your sector of work recently? YG acknowledged:
One of the biggest challenges in the AI space today is the gap between what AI can demonstrate and what businesses can reliably deploy.
AI agents have made incredible progress, but enterprise environments require a much higher standard than a successful demo. A workflow that works 80% of the time is often not enough when it involves sensitive data, customer operations, or mission-critical processes.
The challenge has been shifting the conversation from capability to reliability. Businesses are no longer asking only whether AI can complete a task. They are asking whether it can do so securely, consistently, and at scale.
That challenge has shaped how we build Deck. Instead of focusing only on what models can understand, we focus on the infrastructure required to make those systems dependable in real-world environments. That means handling authentication, permissions, sessions, and the complexity of software systems that were designed for humans rather than AI.
The companies that successfully adopt AI will not just be the ones with the most advanced models. They will be the ones that build the infrastructure needed to make AI trustworthy enough to operate in production.
Evolution Of The Company’s Technology
How has the company’s technology evolved since launching? YG noted:
When we started, we were solving a more specific problem, helping companies access information and automate workflows that were difficult to reach through traditional integrations.
What changed was what we kept seeing across customers. It did not matter the industry or the use case. The same pattern came up every time. Critical workflows trapped inside software that required human interaction, with no API available and no clean way in. That consistency pushed us to think bigger, from solving individual access problems to building a more comprehensive execution layer for AI agents.
The other thing that has changed is scale. Workflow volume has grown more than 300% over the last six months, and the nature of what customers are asking for has shifted with it. A year ago, most conversations were about what was possible. Now they are about what is reliable, secure, and ready for production. That shift has shaped everything about how we build.
Significant Milestones
What have been some of the company’s most significant milestones? YG cited:
One of our most significant milestones was the launch of the new Deck in April of this year. It represented an important evolution of the platform and reflected much of what we had learned from working closely with customers as AI adoption accelerated.
With this updated version of the API, companies are now able to launch workflows significantly faster while achieving success rates of 99%. More importantly, it marked a shift from proving what AI agents could do to helping organizations deploy them with the level of reliability required for production environments.
For me, the milestone reinforced that the market is entering a new phase. Businesses are no longer evaluating AI based solely on capability. They are increasingly focused on execution, reliability, and how quickly they can move from experimentation to real operational value. Seeing customers adopt the new platform validated our belief that the infrastructure layer will play a critical role in the broader adoption of AI.
Customer Success Stories
Can you share any specific customer success stories? YG highlighted:
One example is our work with Evive, a fast-growing healthy food brand that needed better access to retail sales data across its growing network of distributors and retailers.
Their challenge was that the data lived across multiple retailer portals, each with different login flows, authentication requirements, and reporting systems. Since those platforms didn’t offer APIs, their analytics infrastructure was limited by the difficulty of accessing the data itself.
With Deck, Evive was able to automate those workflows. Our agents securely access retailer portals, navigate the interfaces, retrieve reports, and deliver that information directly into their analytics systems.
The result was real-time sales data access across nine retailer portals, more than $100,000 in annual manual work savings, and no ongoing connector maintenance. More broadly, it’s a good example of how we help companies unlock valuable data and workflows trapped inside existing software systems.
Funding/Revenue
Are you able to discuss funding and/or revenue metrics? YG revealed:
We have raised $25 million in funding from investors including Infinity Ventures, Better Tomorrow Ventures, Intact Ventures, Luge Capital, and Golden Ventures, alongside notable angels including Larry Fitzgerald (Pro Football Hall of Fame) and Rahul Mehta (Managing Partner, DST Global).
Total Addressable Market (TAM)
What total addressable market (TAM) size is the company pursuing? YG assessed:
The opportunity for Deck is broad because the challenge we are solving exists anywhere businesses rely on software systems that were not built for AI.
Every industry has critical workflows spread across applications, portals, and legacy systems. Financial services companies, healthcare organizations, utilities, insurance providers, SaaS companies, and operations teams all face the same fundamental challenge. Important work is often trapped inside systems that require human interaction.
We do not view Deck as a solution for a single vertical. We see it as infrastructure that can support AI-driven software execution across the broader software ecosystem. The long-term opportunity is making the existing software world programmable for AI.
Just as previous infrastructure shifts created new categories of products by opening access to previously closed systems, we believe AI agents will require a new layer of infrastructure that allows them to operate across the software landscape.
Differentiation From The Competition
What differentiates the company from its competition? YG affirmed:
The most important thing to understand about Deck is that we are not competing with AI models or agent platforms. We are the execution layer underneath them.
Most teams that eventually come to us start by trying to build this themselves. The prototype works, and then they hit real authentication flows, MFA challenges, bot detection, and interfaces that change without warning. The internal build becomes a maintenance problem nobody signed up for.
What makes Deck different is that we were built specifically for that reality. We operate inside logged-in, UI-based environments, which is the hardest part of automation and where most alternatives break down. Legacy RPA tools work well for stable, structured processes but struggle with real-world variability. Foundation models are exceptional at reasoning but cannot reliably execute across systems without something handling authentication, session state, and edge cases underneath them.
Deck also disappears inside whatever our customers are building. The end user never knows we exist. And rather than producing one-off scripts that need constant maintenance, we convert repeated workflows into reusable, production-grade skills that hold up when interfaces change.
The goal has never been to build another agent application. It has been to make agents actually work inside the software businesses already depend on.
Future Company Goals
What are some of the company’s future goals? YG emphasized:
Our long-term goal is for Deck to become the go-to execution layer that allows AI systems to reliably operate across the software ecosystem. We want to own the category of programmable web infrastructure.
Today, AI systems are increasingly capable of understanding information, analyzing problems, and generating responses. The next major challenge is enabling them to take action in the environments where work actually happens.
We want to continue improving the reliability, security, and scalability of AI execution so companies can confidently deploy agents across increasingly complex workflows.
We also believe the future of AI will be shaped by practical adoption rather than overnight replacement of existing systems. Businesses will continue relying on the software they already use, and the opportunity is making those systems more powerful through AI.
Ultimately, our vision is a world where software becomes more adaptive and where AI helps people accomplish more by handling the complexity of everyday workflows.
Additional Thoughts
Any other topics you would like to discuss? YG concluded:
One of the biggest questions emerging as AI agents become more capable is not just what they can do, but how organizations decide to manage their role inside the business.
A lot of the early conversations around AI adoption have focused on the potential uses and how quickly these systems are improving in performance. But as organizations begin relying on agents for more meaningful workflows, a different set of questions starts to emerge. How should companies define the role of these systems? What level of access makes sense? Where should additional review or oversight be built into the process?
I think organizations should start considering questions before they become urgent. Many companies are already making decisions about how agents interact with their systems and processes, but those decisions are often happening informally. Bringing more structure to those decisions will become increasingly important as agents take on larger responsibilities.
The next major AI breakthroughs will depend on whether organizations can move from simply enabling agents to intentionally defining how those agents operate, what they can access, and where accountability sits.
