There's a category of software that only exists because custom software was too expensive. Flexible project management tools. Configurable CRMs. No-code database builders. Spreadsheets pretending to be apps.
These products don't solve your problem. They give you a toolkit and say: "Build the solution yourself. Inside our platform. With our blocks."
That made sense when the alternative - actual custom software - cost six figures and took six months. When the only way to get something tailored was to hire developers.
That trade-off is dying.
The configuration tax
Every generic tool charges a hidden tax: configuration time.
You sign up for Monday or Airtable or Notion. Day one, you have nothing. An empty workspace. Now you need to build your workflow. Create the boards. Define the fields. Connect the automations. Watch YouTube tutorials. Read documentation. Join the community forum to figure out why your formula doesn't work.
Two weeks later, you have something. It kind of works. It doesn't quite match how your business actually operates, but you've bent your process enough to fit the tool. Good enough.
This is what people have accepted as normal. But it's not normal. It's a compromise forced by economics. You couldn't afford custom software, so you settled for a construction kit.
Custom just got cheap
When a physiotherapist can describe their scheduling needs to an AI and get a working app in a weekend, Calendly with 47 features they don't need stops making sense. When a tattoo studio owner can get a booking-and-consent-form tool built to their exact workflow, they don't need to configure a generic CRM to sort of do the same thing.
The pitch of generic software was always: "We can't build something just for you, but here's something flexible enough that you can make it work."
AI flipped that. Now the answer is: "Actually, we can build something just for you. And it'll take days, not months."
The flexibility that was once the selling point becomes the weakness. Why configure when you can have custom? Why bend your workflow to fit a tool when the tool can be built around your workflow?
Horizontal software was a bridge, not a destination
Think about what Notion actually is. It's a brilliant product. Genuinely well-made. But what problem does it solve?
It solves the problem of not having a developer. It lets non-technical people build internal wikis, project trackers, databases, dashboards - things that would otherwise require custom development. The entire value proposition rests on one assumption: building custom tools is too hard and too expensive for most people.
That assumption held for 20 years. It's crumbling now.
When building a custom internal wiki takes an afternoon with AI, Notion's value proposition weakens. Not to zero - the collaboration features, the ecosystem, the network effects still matter. But the core pitch - "you can build anything here without code" - loses its edge when you can build anything anywhere without code.
The same applies to Airtable, Coda, Monday, ClickUp, and every tool in the "flexible platform" category. They were bridges between what people needed and what they could afford. The bridge is less necessary when the destination got closer.
The unbundling nobody expected
Everyone talked about the unbundling of big software. Slack unbundled email. Figma unbundled Photoshop. Linear unbundled Jira.
But that was unbundling by category. One big tool replaced by a better, more focused tool - still serving millions of users with the same product.
What's happening now is unbundling by customer. Not one tool for all tattoo studios. This tool for this tattoo studio. Not one CRM for all physiotherapists. A CRM that works exactly like this practice runs.
The logical endpoint isn't 10 competitors in each vertical. It's thousands of micro-tools, each built for a specific context, each more useful than any generic alternative because they were built by someone who lives in that context.
What survives
Not everything generic dies. Infrastructure survives. Stripe will be fine - payments are payments. AWS will be fine - servers are servers. Communication tools survive because their value is the network, not the configuration.
What doesn't survive is the middle layer. The "build your own solution inside our platform" category. The tools that exist precisely because custom was too expensive. That's the layer AI compresses.
Some will pivot. Some will become AI-powered themselves - "describe your workflow and we'll configure it for you." That buys time. But it's also an admission: the configuration was never the product. It was the obstacle.
What people actually wanted all along
Generic software was never what people wanted. It was what people settled for.
Nobody dreams of spending two weeks configuring a project management tool. They dream of a tool that just works for their specific situation. That dream was too expensive for 30 years.
It isn't anymore. And every product whose value proposition was "flexible enough to almost fit" is about to learn what happens when "exactly right" becomes cheap.