Software Built For You

For decades, enterprises leaned heavily on IT to build custom software that aligned with their unique needs and processes. One notable example is Walmart in the 1980s, which developed a proprietary inventory management system that revolutionized retail logistics. By investing heavily in a technology project like this, Walmart created a real-time inventory and supply chain tracking system that gave them unprecedented control over stock levels, reduced waste, and optimized restocking. This system became a cornerstone of the company’s competitive advantage, enabling it to scale rapidly and keep costs lower than competitors who relied on more traditional, manual processes and, to this day, helps power Walmart’s position as the largest company by revenue in the United States. 

Bespoke software allowed companies to craft solutions tailored to their competitive advantages, creating differentiated ways of working that competitors couldn’t easily replicate. However, as enterprise technology matured, the complexity and cost of maintaining custom-built systems grew, and as IT teams grew and power structures ensued, they became more isolated from the functions they were meant to support. 

As internet infrastructure matured in the late 1990s, companies began exploring web-based applications to eliminate installation hassles and provide continuous updates directly through a browser. Salesforce, founded in 1999, is widely credited with pioneering the SaaS model by offering CRM (Customer Relationship Management) as a web-based service. Their tagline, "No Software," as paradoxical as it seemed, highlighted the departure from traditional software models and internal IT teams.

While Salesforce weathered the dot-com bubble, its traction reflected a broader trend: companies were increasingly frustrated with large, slow IT teams. In 2006, Amazon Web Services (AWS) launched, allowing startups and enterprises to host applications on Amazon’s infrastructure, dramatically reducing costs and accelerating SaaS adoption. AWS transformed IT with several key benefits:

• Lower Upfront Costs: No need to purchase expensive licenses.
• Scalability: Companies could start small and scale usage as needed.
• Continuous Updates: Software was updated seamlessly without disrupting users.
• Accessibility: Applications could be accessed from any device with an internet connection.

AWS was, at first, particularly transformative for startups, cutting the cost and complexity of launching technology by offering storage and compute as services. Larger enterprises also benefited, as AWS enabled them to modernize legacy systems, experiment with new applications, and scale infrastructure without significant capital expenditure. From the 2010s onward, companies like Dropbox, Slack, Zoom, and Google expanded SaaS into productivity, communication, and collaboration, and consultants like Accenture, Deloitte, and others charged hefty fees in moving enterprises off traditional IT systems and teams under the investment banner that would rule corporate boardrooms for well over a decade, digital transformation.

When we started Percolate in 2011, fully leveraging SaaS apps and cloud infrastructure wasn’t yet the default in most enterprises. But by 2015, that had changed. Even large enterprises transitioned to SaaS for ERP, HR, and security tools. Microsoft (Office 365), Adobe (Creative Cloud), and even the largest software company holdout, Oracle, migrated their products into SaaS offerings, while SaaS providers increasingly integrated with one another, creating ecosystems that facilitated interconnected workflows.

By the 2020s, SaaS had overtaken IT around the world, and the largest enterprise SaaS companies had become some of the world’s most valuable businesses by market cap:

• Salesforce: $300B+
• ServiceNow: $200B+
• Palo Alto Networks $100B+
• CrowdStrike: $50B+
• Workday: $50B+

March 1st, 2023: The Script is Flipped on SaaS

On March 23, 2023, OpenAI released the first ChatGPT plugins API, allowing developers to integrate ChatGPT’s capabilities directly into their applications and services. This marked a new turning point. Noah, my co-founder at Alephic, described the moment he saw the announcement—he canceled his meetings immediately to experiment with the API. Unlike traditional APIs, OpenAI’s approach flipped the paradigm. You simply provided data rather than adapting to complex API specifications, and the system handled the rest.

This experience, as Noah recounted, was one of the most counterintuitive and mind-blowing shifts he had ever seen in his 20 years working in technology. He also started to think about all the implications this would have as AI not only changed how you would interact with an API, it would also change the potential for the type of software you could build. Noah documented his experience a few times with Building ChatGPT Plugins and AI as a Fuzzy Interface.  It has been a year and a half since the API release, and I can confirm that his mind is still blown, unpacking all the different implications AI has on building software in the future. 

So, let’s go back to the beginning to look at where we are going. 

Forty years ago, enterprises built custom software—it was expensive and slow but differentiated. Then came the SaaS revolution, which stripped away IT bottlenecks at the cost of competitive software advantages. Enterprises conformed to SaaS products rather than the other way around. Meanwhile, tech giants like Apple, Google, Microsoft, Meta, and Amazon retained their advantage by continuing to build proprietary software, becoming trillion-dollar companies in the process.

The release of OpenAI’s Plugin API—or at least what it represents—signals a shift back towards building custom software. Leveraging AI, coupled with an enterprise's ability to write code and harness its collective internal expertise, allows organizations to build systems tailored to their unique DNA. This shift undermines the traditional SaaS model, which forces companies to fit into rigid templates that allow the generalized software to service as many companies as possible. Now, enterprises can develop AI-driven platforms that adapt to them.

This is the most exciting era for enterprise technology, as AI democratizes access to powerful tools once reserved for only tech giants who had the compute and engineering talent to build everything custom. Enterprises can now harness AI to create software systems that evolve with their needs, positioning technology as a core driver of competitive advantage. The days of multi-year software implementations that also diminish your differentiation are fading. With AI, technology conforms to your enterprise, not the reverse. SaaS had a great run, but the future increasingly belongs to companies that build their own software. Powered by AI, of course.