AI Has Already Changed How Coders Work – Now It Is Coming For The Rest Of Us

For the last few years, the AI productivity story has been essentially a developer story – tools like Claude Code transformed how engineers write, review and ship code, Cursor became a $29 billion company, and GitHub Copilot went from curiosity to standard kit.

The developer market got transformed, arguably faster than anyone expected, and the AI labs that powered that transformation got very rich in the process. The problem is, the developer market just isn’t that big.

There are roughly 27 million software developers in the world. Meanwhile, there are roughly 3.5 billion knowledge workers – the math isn’t subtle. The market that holds greater potential, the one that AI labs, enterprise software companies and a generation of startups are now positioning to capture, is the one that has never written a line of code.

That shift is already underway, and it’s going to reshape the SaaS landscape far more dramatically than the coding revolution did.

The Developer Market Was Just The Warm-Up Act

The pattern in this is worth understanding, because it explains a lot about what’s coming.

AI tools went to developers first for a simple reason: developers could evaluate them. If a coding assistant wrote bad code, an engineer could flag it immediately. That efficient feedback loop meant early AI tools could improve fast, get adopted fast and generate revenue fast. It was a brilliant front line: high willingness to pay, high technical tolerance, high word-of-mouth velocity.

But no serious AI lab was ever planning to stop there. The tools that emerged from that developer-first phase – agents that can plan multi-step tasks, browse the web, write and execute code, manage files, send communications – are now being repackaged for a completely different audience. They’re becoming personal concierges, executive assistants, document processors, customer service tools, presentation generators and inbox managers. The target audience is, in the most literal sense, everyone.

Anthropic, OpenAI, Google and others are all moving in this direction simultaneously. The race to own the non-coder market is in full swing.

A Plot Twist Saas Founders Were Not Ready For

If you’re building a SaaS product in the productivity space right now, this is the part worth contemplating.

The value proposition of most productivity SaaS tools is some version of: “we make a specific task faster, easier or more organised,” – a perfectly good value proposition. It’s also one that an AI agent running in the background can now replicate, sometimes better, for a fraction of the cost.

The incoming wave isn’t AI layered onto existing products as a feature nobody had on their wishlist. It’s AI as a wholesale replacement for the category. An AI that manages your inbox doesn’t just add a smart filter to your email client. It makes you question whether you need the email client at all.

That’s not hyperbole, it’s the direction multiple major labs are explicitly building toward. The data on how businesses are actually using AI suggests most companies haven’t gotten close to understanding what that means for their stack yet.

What The Non-Coder Market Looks Like

The non-coder opportunity breaks into a few distinct zones, and they’re not all equally accessible.

The first is personal productivity. Scheduling, email triage, document drafting, presentation creation, research summarisation, this is the territory that tools like Notion AI, Microsoft Copilot and Google Workspace AI are already occupying. It’s a massive market and the competition is fierce, which means the realistic opportunity for new startups is increasingly in the niches, specific industries, specific workflows, specific user types who need something the horizontal platforms won’t build.

The second is professional service automation. High-value, specialised tasks like legal document review, financial report drafting, medical note-taking and marketing copy generation all sit in this zone – work where AI agents can replace significant chunks of professional workflow and willingness to pay is correspondingly high.

The third, and most transformative, is agentic automation. Not just AI that assists with a task, but AI that completes tasks autonomously: booking, ordering, filing, communicating, deciding. This is still early, but it’s the direction the biggest AI labs are investing in most aggressively, and it’s the zone that will create the most disruption to existing SaaS categories.

Should Founders Be Excited Or Terrified?

Both, honestly, but probably more excited than the current level of anxiety in the room suggests.

The non-coder market being enormous doesn’t mean it’s going to be captured by a handful of large labs – it’s too fragmented, too specific and too relationship-dependent for that. The workflows that matter most to a law firm in Birmingham aren’t the same as the ones that matter to a logistics company in Rotterdam or a marketing agency in Morocco.

General-purpose AI agents are going to be excellent at general-purpose tasks. The specific, high-stakes, deeply embedded workflows are still going to need someone who understands the category. The founder opportunity sits in that gap: building something so specific and so trusted in a particular context that the horizontal platforms can’t replicate it even if they try.

The developer market taught us that AI doesn’t necessarily kill existing tools, it often raises the floor for what users expect. The next unicorn probably isn’t a general AI assistant. It’s probably a deeply vertical product that happens to have AI at its core, built by someone who understands a specific market better than any lab ever will.

The non-coder wave is coming. Founders who move now, before the horizontal platforms fully arrive, are the ones who’ll have the best chance of owning something worth owning.