By Natalie Johnson
The software industry spent 20 years perfecting the art of selling humans tools to do human work. Then, quietly, over a single holiday week in late December 2025, that premise became obsolete. Daniel Saks, CEO of Landbase and co-founder of AppDirect, was on a boat in Indonesia when it happened; seven days without connectivity, somewhere between Christmas and the New Year. When he reconnected, the messages waiting for him told a story he had not expected so soon: the engineers and operators he respected most had, in a matter of weeks, written more code than they had in their entire careers combined.
The founding premise of Landbase, that people should not have to work for their software, but their software should work for them, had just become reality. And almost nobody had noticed yet. “Instead of a human telling software what to do,” Saks said, “the software is doing the work for the human.”
The SaaS Apocalypse Is the Wrong Frame
By early Q1 2026, the markets had a name for what Saks had sensed on that boat. They called it the ‘SaaS apocalypse’. He finds the framing misleading, not because the disruption is overstated, but because it misidentifies what is actually being destroyed. What is dying is a pricing model, not an industry. Any software as a service (SaaS) company that built its revenue architecture around human seats is facing structural collapse. Fewer humans will be employed to tell software what to do, which breaks the predictable growth curves that defined SaaS valuations for a decade. But software itself is not contracting. It is expanding at a rate that makes the previous era look tentative.
“More compute is going to be used to create more software to drive more productivity than ever before,” Saks argues. “Agents will become the primary user, and agents will likely consume more compute and demand than ever before.” The result is not an apocalypse. It is a sorting between the SaaS companies that figure out how to harness their data and intelligence in an agent-first world, and those that stagnate with business models designed for a dynamic that no longer exists.
A Furniture Store in Niagara Falls
Saks has seen this kind of sorting before, on a smaller and more personal scale. His family ran a century-old furniture store on Main Street in Niagara Falls, Canada. It closed in 2009, the year IKEA and the first wave of digital retail finished off most of the small shops on that block. “My family’s furniture store may have existed if they had sold online or if they had acquired others,” he reflects. “But they didn’t. They just stagnated the old way, and it closed.” The lesson he carries into the AI era is that proactive self-disruption is not a strategic option. It is a survival requirement.
He points to Intercom as the clearest modern case study. The company spun up a new group called Fin, hired against its own legacy business, and built what is now a fast-growing AI-native company that has likely eclipsed Intercom’s traditional market cap. His former company, AppDirect, approaching a billion in gross annual recurring revenue (ARR), has launched Devs AI, a product Saks believes carries a distribution and data advantage that most traditional SaaS players cannot replicate. Salesforce is pushing hard through Agentforce. Figma’s stock has collapsed roughly 80% as it searches for an AI-native footing. The transition is happening in real time, and most organizations are still deciding whether to take it seriously.
Drink Your Own Champagne
Saks runs Landbase on the AI it builds. That includes deploying it within Claude Code to automate significant portions of his own workflow as CEO, a practice he describes as completely transformative for both his productivity and his understanding of why Landbase will be differentiated from competitors.
His advice to every professional, regardless of role or technical background, is to download Claude Code and start automating your own workflow. “Those who started a month earlier already had a huge head start,” he notes. “That’s how stark it is.” Landbase’s three internal values – own it, automate it, elevate it – capture the operating philosophy that he thinks every professional and organization needs to internalize.
What Executives Are Still Missing
The question most leaders are asking is which AI tools to buy. Saks thinks that is the wrong question. The models are already powerful enough. The constraint is not capability, it is context. “For a CEO to be prepared, they need to ensure that their own internal data is structured in a way that agents can read and make decisions on,” he argues. His warning to organizations still treating ChatGPT as their primary AI interface is precise: “If you’re just using ChatGPT today, you’re missing out on 98% of the value of AI.” The real power comes from giving AI memory, data, and organizational context, and then deploying a code provider like Claude Code or OpenAI Codex to act on it.
The day SaaS stopped needing humans was not the end of software. It was the beginning of an era where the companies and individuals that understand what that shift actually means will build things that were previously impossible, and those that do not will spend the next decade wondering what happened.
Follow Daniel Saks on LinkedIn for more on the future of AI-native software, agentic go-to-market, and building in the post-SaaS era.
