The Human Interface Disruption
September 2026
Written by: Mazy Dar
SaaS·poc·a·lypse (/sɑːs-pə-ˈka-lɪps/) n. A 2026 fear that AI coding tools would let companies build their own software and drop paid subscriptions, sparking a software stock selloff that later proved overblown.
Earlier this year, the market briefly convinced itself that AI coding tools would let enterprises build their own software and cancel their SaaS subscriptions. Software stocks sold off accordingly. At Dreamforce a few weeks ago, Marc Benioff called the whole idea "crazy nonsense."
Of course he's talking his own book, but he's also right.
I'm not foolish enough to claim enterprises won't try. Banks in particular have a long history of building things that have nothing to do with their core competency: email systems, chat platforms, even their own operating systems. So it's not a question of whether someone tries this. It's a question of whether it works, and for a system of record, the answer is almost always no.
Writing the new system is the easy part. Migrating to it and maintaining it is a different story entirely. A system of record carries compliance weight at any enterprise of meaningful size: audit trails an examiner can ask for years later, access controls tied to regulatory obligations, data-handling rules that vary by country, state or industry, and integrations with a dozen other systems that all assume the old plumbing is still there. Replace that plumbing yourself and you've taken on the liability your vendor used to carry, with none of the track record to back it up. The pain and risk of a project like that far outweigh whatever you'd save by no longer paying your SaaS vendor. And once you've done it, congratulations: you now own an entire system that has nothing to do with your core business, and it's your job to maintain it forever.
So why is the idea so tempting? Because most enterprises pay a fortune for these platforms, and their people still dislike using them. When AI makes building software look easy, it's natural to wonder why you're paying so much for something nobody enjoys. But that frustration is aimed at the wrong target. The platform isn't the problem. The screen is.
Every major SaaS vendor is solving an impossible problem: one user interface that has to work for every customer, every industry, and every role. Generally means it's never quite right for anyone. The vendors know it, which is why they give you filters, dashboards, custom fields, and drag-and-drop builders to bend the screen into shape. Most of it is a pain to use, and what you end up with is still a compromise: a view that's a little more yours, and a lot more work to keep running. The value you're paying for is real. It's just locked behind an interface that was built for everyone, which means it wasn't built for you.
This is the layer AI is actually good at replacing, and the layer it's safe to replace. The APIs already exist. The data's already exposed, because the vendor needs it for their own integrations, their own automation, their own mobile app. Build a new interface on top of that API and you never touch the database, the audit trail, or the system of record. You touch the part people actually use and you make the system far more useful than it’s ever been. You won’t save the subscription fee doing this, but you will unlock enormous value from the platform you're already paying for.
And here's the best part: you don't need a dev team to build it, and you don't need to live with whatever the vendor's builder lets you drag and drop. You describe how you want it to work, in plain language, and you get an interface that's actually yours. Not a template with your logo on it; something tuned to how your organization actually operates, customized for the specific persona using it. Your onboarding team sees a different view than your support team, because they're doing different jobs and staring at the same underlying data for completely different reasons.
Replacing the interface unlocks another important and related opportunity. It's not just that any single app's interface is bad; it's that most workflows span multiple systems, and your SaaS vendors don’t build their interfaces to support that. A salesperson pulls up a client and wants everything relevant to that relationship in one place: the Salesforce record, the last Outlook threads, the Gong call notes, the open Jira tickets, the Zendesk history — not five tabs and five mental models stitched together in their head. Every team wants this in some form. Almost nobody has it, because these systems were built in isolation from each other, not alongside each other.
That's what got lost in the SaaSpocalypse conversation. Replacing a SaaS platform is a monumental task, but replacing its dated, inflexible interface is now easy with AI and it’s the key to unlocking productivity across all the systems that run your business.
Benioff is right. The major SaaS platforms that are your systems of record aren’t going away anytime soon. But their human interfaces are ripe for disruption. You can build an interface on top of your existing Salesforce or Dynamics that's actually tuned to how a specific team works; not one generic screen trying to be everything to everyone, but something shaped around the persona using it. A ticketing system that shows the issues for a particular customer without being asked. A workspace where the systems someone needs are all sitting next to each other and contextually connected, instead of being isolated across five tabs.
This is the opportunity we built HERE Studio to help unlock. Studio is the AI app builder inside HERE, our governed workspace. Studio lets anyone describe the interface they want in plain language and have it built with their own company’s approved AI on top of the SaaS platforms they already run, with permissions, compliance, and branding inherited automatically, and running as part of one unified workspace instead of another disconnected tab. If this resonates, I'd love to show you what we've built.