Mobile AI 2026: What Users Actually Want
AI features in mobile apps everywhere. But usage data tells a different story than what the hype suggests. Here's what people are actually using and paying for.
Enterprise mobile apps are famously unusable. Here are the five most common UX failures I see, and why they're easy to fix.
Abhi Asok
Founder & CEO, Arvension Technologies
I asked a CTO about their mobile app. His response: "Nobody uses it. Field teams prefer taking photos and emailing them to the office."
The app exists. It's full-featured. It connects to the backend. And it's so bad that field teams prefer 2005 workflows instead.
This is what enterprise mobile looks like in 2017: technically possible, functionally complete, and practically unusable.
I've audited dozens of enterprise mobile apps. I see the same mistakes over and over. They're not technical mistakes. They're UX mistakes. And they're almost always fixable with no architecture changes. The teams just didn't prioritize the experience.
The most common mistake: take the desktop web app, put it on a phone screen, and call it mobile.
Lists that scroll horizontally. Buttons that are 2mm wide. Forms with 15 fields. Nested menus four levels deep. Every interaction that works on a 24-inch monitor creates a usability nightmare on a 5-inch phone.
I was looking at an inventory app. The search screen had five dropdown filters. On mobile, they stack vertically, forcing users to scroll past all of them to see results. Every use of the app requires scrolling through filters first.
Fix: mobile-first design. What's the core task? Search inventory. Do that first. Hide the filters unless explicitly requested. Let the user search with one tap. Then, optionally, filter.
This requires saying no. Enterprise teams hate saying no. So they build a smaller version of everything, and the result is worse than everything.
Field teams don't have consistent connectivity. They're in warehouses, manufacturing facilities, rural areas. But most enterprise apps assume internet.
The field team starts the day with connectivity, opens the app, and loads their tasks. They go offline. They complete a task, try to submit, and it fails. They either give up or screenshot it.
I built an app for a logistics company requiring internet connectivity for delivery lists. When the driver lost signal, the app was useless. The company was paying for mobile apps to replace clipboards. The mobile app was actually worse because clipboards work offline.
Fix: service workers. Local storage. Sync when connectivity returns. This is technically straightforward in 2017. But it requires architecture changes upfront. Most teams add it later, if at all.
Enterprise workflows are complex. Typically they require context. You're reviewing an order, which pulls from a customer record, which references contracts, which pulls from a pricing table. Everything is connected.
On desktop, you can have multiple windows open. You can scroll. You can navigate hierarchically. On a phone, that depth of information becomes a navigation maze.
I saw an order management app requiring tapping through six different screens to review an order. On desktop, you see everything at once. On mobile, it's a scavenger hunt.
Fix: ruthless scoping. What's the core decision the user needs to make right now? What information is necessary? Everything else is secondary. Push it into a detail view or remove it.
This usually requires involving field teams in design. "What do you actually need on your phone?" Usually, it's 20% of what the desktop app shows. That 20% changes your priority.
Related to offline: the app stores nothing locally. Every screen requires a fresh network request. Scroll through a list, select an item, wait for data. Navigate back, it refreshes the list from the server again.
This creates a choppy experience. It's slow. It drains battery. Dependent on network quality.
I tested an app from a major software vendor. Load 100 items: 8 seconds over cellular. Select item: 3 more seconds. Go back: 8 more seconds. Five interactions takes 20 seconds.
Fix: cache everything. Load a list, store it locally. Display instantly next time. Refresh in the background. Field teams tolerate data that's five minutes old if it loads instantly. They won't tolerate waiting 8 seconds per interaction.
Enterprise apps serve multiple users with different needs. Managers, field workers, supervisors, back-office staff. But most apps show everyone the same interface.
A warehouse manager doesn't need the same view as a warehouse worker. But the app shows them the same thing, just with permissions hiding some buttons.
I tested a supply chain app. A planner, buyer, and warehouse manager all saw the same screen despite different jobs and needs.
Fix: role-based interfaces. When users log in, show them the interface for their role. Remove the noise. Show their decisions. Everything else is secondary.
All five mistakes have the same root cause: desktop-first thinking. Teams build the desktop app, then adapt it for mobile. Mobile becomes a secondary platform.
Companies winning in enterprise mobile do the opposite. They start with mobile. They understand what's actually necessary for field users. They optimize for offline. They make every interaction count.
It's not harder. It's different. It requires thinking about what users actually need rather than replicating desktop functionality.
But here's the payoff: when mobile is genuinely usable, field teams adopt it. They stop using clipboards. The app becomes infrastructure instead of theater.
The question for 2018 isn't whether you should build mobile apps. It's whether you're building them with field users in mind or just adapting your desktop experience and calling it done. The difference is everything.
AI features in mobile apps everywhere. But usage data tells a different story than what the hype suggests. Here's what people are actually using and paying for.
Nine years building mobile apps. The native vs cross-platform debate isn't settled by hype. Here's what the data actually shows and what it means for your next app.
On-device AI models (Apple Intelligence, Google Gemini Nano) make offline-first AI possible. Real patterns for building features that work without internet.