Why I joined Everest

I’ve lived in the San Francisco Bay Area almost my entire life, and I’ve witnessed first-hand, the tech-led disruptions from the advent of personal computing, dot-com, mobile, and now the AI revolution.  Aside from my time at JPMorgan, I have been directly immersed in the tech industry from starting a company in the dot-com boom, to pivoting the company into enterprise software, to helping SAP transform its user experience to an award-winning consumer-like experience.  The AI wave feels bigger than all of these previous disruptions.  I did not want to be just a consumer of this.  It was time to re-enter the tech industry and be a part of an organization that could help define it.

Earlier last year,  I visited my old colleague Franz Faerber in Heidelberg to understand what he and the team at Everest had been building and though I’d initially thought of it as an ERP, I realized that it’s actually much larger.  It was positioned as AI-native, ERP built to leverage all. However, when I learned that Everest was also productizing the AI platform that they used to allow companies to design, specify, generate, deploy, and maintain applications, another term came to mind - “Enterprise-native AI.”

At any large enterprise the great preponderance of IT budget is tied up “keeping the lights on.”  This includes recurring fees tied up in SaaS licenses and servicing old tech debt. Everest, I have come to realize, could help businesses unlock the topics  of tech debt, underperforming software licenses, and addressing opportunities to deliver business value regardless of what core ERP it’s running on. In this article, I share my journey to joining and the paradigm shift this product represents. 

From Design to Enterprise AI

My career started as a co-founder in an enterprise software company that my business partners and I sold to a competitor.  We were close partners with SAP, which I joined. That was a wonderful chapter in my career. SAP’s founder and chairman Hasso Plattner had gotten really into the topic of design thinking and through his investment established  the Hasso Plattner Institute of  Design at Stanford. I was among the first external hires for a newly created design team situated in the office of the CEO and over 13 years there, I became SAP’s first chief design officer.

The business value of design was never just aesthetics, rather it helped businesses fundamentally understand the needs of their customers better and helped focus on what problem mattered most.  I have often said that true innovation lies at the intersection of creativity and execution.  Execution is dependent on the capacity of an organization’s capacity to solve problems.  On the other hand, creativity is the ability to find the right problem to solve which is fundamentally an output of a design process.  What differentiated design was a focus on the who, the why, and the what before jumping directly to the how.

I can’t help but notice a parallel between the design process and the emergence of generative AI technologies in the workplace.  In 2026, AI has progressed beyond a natural language interface to answer questions, to agents that complete tasks.  In our previous example, the execution becomes automated and much more efficient.  However, the creativity part still relies on the ability to find the right problem to solve - discovered through questions, queries, and prompts.

Code generation can quickly exacerbate tech-debt and maintenance costs

One of the most promising values of AI agents has been in the area of developer efficiency, that have rapidly advanced from code completion, automated documentation, test coverage, to code generation tools such as Lovable, Cursor, GitHub CoPilot, Claude Code and others.  However, I see that in an industry rushing to commercialize AI, especially in the enterprise. Hype has preceded use. The result is a lot of unusable code and unmaintainable or insecure applications that can’t actually make it into production. Now the amount of tech debt is worse than ever.

The trouble here is not a lack of inventiveness by those LLM companies, but a lack of attentiveness to finding the right problem to solve.  It’s not just about code generation - rather it’s about delivering value to the business into production in a scalable and maintainable way. The frameworks to make this happen already exist. It’s called the software development lifecycle, it’s plenty-well defined, and the software development part is just a fraction of the overall process. Yet AI tools don’t abide by it. They don’t allow users to collaborate, roll back changes, branch and merge, or co-define specifications, and deliver solutions into production that are secure, robust, and scalable. 

AiSpecify by Everest is the first-ever enterprise-native AI solution

Everest, I realize, is that. The founders started by building a fully functional ERP with everything a mid-to-large enterprise needs to run, but more importantly, they built an AI solution to build new features on top of it. Picture your current ERP and the tech stack that runs on top of it. Now imagine you don’t have to pay for all custom development or subscription costs to all those bolt-on solutions because you can just prompt to produce them.

It is more complicated than that, of course, and that is what I believe is the true genius of Everest. Everest’s AI feature-producing tool, AiSpecify, starts by helping teams design requirements and  specifications together. Then, it generates the code and applications, moving them through increasingly realistic environments through testing, deployment, and support. If something needs changing, you edit the specification and it regenerates the code and all documentation. 

It is worth pausing to consider what a profound difference this is. It’s the power of AI-generated code, locked down and usable in enterprise workflows specific to your organization. It’s made for the CIO overseeing 1,000 or 10,000 developers. It’s for product and customer success organizations. It’s for more than just prototypes—it finally connects experimentation to production.

Code is easy, making it enterprise ready is hard

Coding a new feature or functionality used to be hard, but that was never the actual thing blocking large enterprises from innovating. Otherwise vibe coding would have unleashed a wave of product and customer experience innovation. It was creating applications and features that were secure, governable, auditable, and reliable.  But now, I believe that enterprise-native AI applications like Everest’s AiSpecify have closed that gap. By starting on a secure foundation and following the software development lifecycle, they’ve created a product that can help organizations cut into  the maintenance burden companies pay on tech debt and the SaaS solutions surrounding core systems of record. And it was a big enough opportunity that I wanted to be here with this team bringing it to market.

Finally, it’s not just what you do, but who you work with.  I’m genuinely excited to work with some of my esteemed, former colleagues again to help define the next generation of enterprise solutions.  While the promise of AI is vast, our solutions remain focused on the empowerment of people to achieve superior business outcomes.

If this topic interests you, I’d love to hear from you at sam.yen@everest-systems.com.