In the age of AI, startups need to do two things exceptionally well. First, they need to solve problems, not sell software. As Sequoia argued earlier this year, the most valuable companies of the next decade may look less like traditional software businesses and more like services firms powered by software. Customers don't want another tool; they want an outcome.

Second, they need to move incredibly fast. The underlying technology is improving at a rate unimaginable just a few years ago and competitive advantages are increasingly short-lived. The market is constantly shifting beneath everyone’s feet.

The resulting complexity: The moment a company decides to own an outcome rather than simply provide software, it inherits every problem standing between the customer and their success. The winners will be the companies that can continuously turn new capabilities and tech into customer outcomes and do so faster than everyone else. The key is execution. Or put  less delicately, the ability to Get Shit Done. 

This is the philosophy we're bringing to eCommerce scale at ZyG. The desired outcome is simple: A founder or company with a great product and early traction wants to reach scale in the U.S. market. But getting scale remains elusive for most.

That’s because eCom scale is the result of dozens of interconnected capabilities working together. It requires building high-converting storefronts, generating compelling creative, running effective acquisition campaigns, creating organic growth loops, improving retention, supporting customers, optimizing logistics, forecasting demand, and continuously identifying the next constraint on growth. It also means building the data layer that connects every part of the business, developing predictive models, and deploying a network of AI agents capable of executing work across the funnel. 

ZyG is AI first - without AI it would be impossible to solve the complexity that is eCom scale. But while the technology matters, it is not the outcome. The outcome is scale. And success comes from bringing all the capabilities necessary, simultaneously and integrating them into an outcome, whether that solution involves software, automation, operational expertise, a human in the loop, or all of the above.  

The History of Getting Shit Down

This is not a new pattern. Every major technological shift eventually reaches a point where invention stops being the bottleneck and execution becomes the differentiator.

The railroad boom of the nineteenth century was not won by the companies with the best locomotives. It was won by those that could lay tracks, coordinate supply chains, finance expansion, manage operations across geographies, and turn a promising technology into a functioning network. The automotive industry followed a similar path. Building a car was an engineering challenge. Building millions of cars, distributing them globally, servicing them, financing them, and continuously improving production was an execution challenge.

The same story played out during the rise of enterprise software. The winners were not simply the companies with the best code. They were the companies that could implement systems inside large organizations, redesign workflows, train employees, and ensure customers achieved measurable business outcomes. The software mattered, but the work required to make the software valuable mattered just as much.

Back to the Present

AI is pushing technology companies into the next phase of this evolution. Enterprise software helped companies move beyond simply selling tools and the best software companies learned that success depended not just on shipping code, but on helping customers implement it, adopt it, and achieve measurable results. But what was once a competitive advantage is quickly becoming the minimum requirement.

In the age of AI, customers are increasingly coming to expect an outcome. But, somewhat paradoxically, despite the seeming magic of AI - from writing a Shakespearean poem in seconds to doing what was once complex coding - reaching an outcome remains extremely rare. That’s because AI is an incredibly important piece of the puzzle, but it is not a magic bullet. There is no prompt that will allow your product to reach scale, for example. It necessitates a combination of deep AI, modeling data expertise, the agility to evolve the tech stack with the shifting technological changes, and the ability to use (human) judgement on what is needed to close all of the loops. 

The last twenty years of software were about helping people do work. The next 20 years will be about doing the work for them. The winners won’t be the companies with the smartest models or the prettiest interfaces - they’ll be the companies that can consistently turn technology into outcomes. In other words, the companies that can get the most shit done.