Insights on Mobile Apps, AI, and Emerging Technology
Short perspectives on the technologies shaping digital products — from mobile apps and AI to cloud platforms, connected devices, and data-driven systems. Read more in-depth articles.
Building or modernizing a digital product? Underlabs helps turn ideas, workflows, and platforms into reliable software.
Discuss your project
-
AI Agents
Software started doing more than waiting for clicks. AI agents pushed apps toward planning, acting, checking, and completing multi-step tasks. The promise is attractive: book, compare, summarize, file, answer, escalate, and repeat. The danger is also obvious: an agent with bad context can make bad decisions faster than a human can catch them. The useful future is not fully autonomous software everywhere. It is supervised automation, with clear permissions, good logs, and humans still holding the steering wheel. See how supervised AI agents fit real workflows.
-
Android On-Device AI
Android's AI stack moved further onto the device. Gemini Nano lets developers build generative AI experiences that can work without sending every request to the cloud. That changes the economics and privacy model of AI features. A mobile app can summarize, classify, suggest, or rewrite without needing a constant network call. The result is not that cloud AI disappears. The result is a new split: small intelligence on-device, heavier reasoning in the cloud, and better products deciding carefully between the two. Read why AI value sits in the application layer.
-
Agentic Mobile Apps
Mobile apps are becoming less like screens and more like helpers. Apple's newer Foundation Models work points toward agentic experiences: apps that can understand context, search semantically, use models, and take structured actions through system frameworks. This is a major change for product design. The old model was buttons and forms. The new model is intent, context, permission, and result. The interface is still there, but the app is expected to understand more of what the user is trying to accomplish. See how supervised AI agents fit real workflows.
-
On-Device AI
AI started moving closer to the phone. Apple's Foundation Models framework gave developers access to on-device language models for privacy-preserving, offline, low-cost intelligence features. That matters because not every AI app feature needs a cloud call, a token bill, or a server round trip. For many apps, the best AI will be small, local, fast, and boring in the right way: rewriting text, classifying content, extracting details, suggesting next steps, and helping users finish tasks. Read why AI value sits in the application layer.
-
AI Inside Normal Apps
The AI app stopped being the only model. Instead of building a chatbot and calling it a product, better teams started putting intelligence inside the existing workflow. A storage app could identify objects. A parking app could read signs. A mediation app could summarize positions and propose next steps. The lesson was simple: AI is strongest when it is attached to a real job. The magic is not the chat window. The magic is removing the work the user did not want to do. See putting AI inside existing workflows.
-
Mini Apps and Super Apps
Small apps inside bigger apps became harder to ignore. Apple introduced a program for mini app developers, reducing commission for certain self-contained HTML5 and JavaScript experiences inside host apps. The idea is old in parts of Asia, especially around super apps, but it keeps returning in new forms. For developers, it raises a practical question: should every product be a full app, or should some experiences live inside a larger surface where the user already spends time?
-
Alternative App Distribution
The App Store wall started to crack, at least in Europe. To comply with the Digital Markets Act, Apple announced changes allowing alternative app marketplaces, alternative browser engines, new payment options, and hundreds of new APIs for EU developers. This did not mean the old App Store model disappeared. It meant distribution became more political, more regional, and more complex. Developers now had to think not only about features, but about where and how an app is allowed to reach users. Read about when developers can challenge App Store rulings.
-
GPT-4o
AI became more conversational, visual, and immediate. GPT-4o pushed text, image, and audio into a faster real-time model. That changed product expectations again. Users no longer wanted to only type into a box. They wanted to speak, show a photo, ask a follow-up, and get useful answers in context. For mobile apps, this made the camera, microphone, and chat interface feel like parts of the same product surface.
-
Private Cloud Compute
Apple drew a line around private AI. With Apple Intelligence and Private Cloud Compute, the message was that some AI should happen on-device, and heavier requests should be processed in a way that does not turn personal data into a training buffet. Whether every detail works as promised is a separate question. The important shift is that privacy became part of the AI architecture discussion, not just a line in a policy page.
-
GPT-4
AI became useful enough to put inside real workflows. GPT-4 improved reasoning, coding, summarization, and multimodal understanding compared with earlier models. It was still not perfect, and it could still be confidently wrong, but the quality jump changed expectations. Users began to expect apps to help them write, decide, search, compare, explain, and automate. For software teams, AI became less of a feature category and more of a new layer inside many products.
-
Web Push on iOS
Progressive web apps gained one of the features they had been missing on iPhone: push notifications. With iOS 16.4, web apps added to the Home Screen could send web push notifications. This did not suddenly make PWAs equal to native apps, but it narrowed the gap for certain products. For internal tools, lightweight SaaS dashboards, and simple customer portals, the question became more interesting: does this need to be in the App Store, or can the web carry enough of the experience?
-
Spatial Computing
Apple put a new name on mixed reality: spatial computing. Vision Pro was not a mass-market device at launch, but it reframed AR and VR around workspaces, media, design, training, and 3D interfaces. The important part was not the headset itself. It was the idea that apps may eventually move beyond flat rectangles on glass. For developers, it opened a new question: what does software look like when the room becomes part of the interface? See our deeper look at Vision Pro app design.
-
SwiftUI Matures
SwiftUI started moving from interesting to serious. With better navigation, charts, layouts, and platform support, Apple's declarative UI framework became harder to ignore. It still did not replace every UIKit use case, especially in large existing apps, but it became a stronger option for new screens and modern interfaces. The bigger shift was mental: developers were no longer describing every step of the interface. They were describing the state, and letting the UI follow. Apple introduced Swift Charts and other major SwiftUI updates at WWDC 2022.
-
The Subscription App Era
The one-time paid app became the exception. More apps moved toward subscriptions, usage limits, credits, trials, and freemium models. Some of it made sense: servers, AI, storage, support, and ongoing app development cost real money. Some of it became abusive: simple utilities asking for monthly rent. The lesson for developers was clear. A subscription is not a pricing trick. It has to match recurring value, or users will feel it immediately. See what belongs in version 1 before monetization and app development cost in Canada.
-
ChatGPT
AI stopped feeling like a research demo. ChatGPT made large language models understandable to normal users, founders, students, developers, and businesses. The interface was simple: type, ask, refine, repeat. The impact was not that AI could magically replace software teams. The impact was that every product team now had to ask where language, reasoning, summarization, classification, and automation could improve the user experience. OpenAI launched ChatGPT on November 30, 2022. Read what happens when AI builds faster than QA keeps up, or use our AI Launch Inspector before shipping.
-
App Tracking Transparency
The ad-supported app economy hit a wall. App Tracking Transparency required apps to ask permission before tracking users across other companies' apps and websites. Many users said no. That forced mobile businesses to rethink attribution, analytics, retargeting, and growth. It also pushed better first-party data, clearer onboarding, and stronger product value. In short, developers could no longer assume that invisible tracking would carry the business model. Apple had announced the permission-based tracking requirement in early 2021.
-
StoreKit 2
In-app purchases became less painful to build properly. StoreKit 2 gave developers cleaner Swift-native APIs for subscriptions, purchases, transaction history, entitlements, refunds, and customer status. That sounds like plumbing, but plumbing is where a lot of app businesses leak money. A subscription app is not just a paywall. It needs receipt logic, renewal handling, restore flows, edge cases, server validation, and clear customer support. StoreKit 2 made that world a bit less fragile.
-
Remote Work Software
Work moved into the app. After the first shock of remote work, companies started asking for better internal tools, dashboards, mobile approvals, chat workflows, client portals, and automation. The demand was no longer "can we work online?" It became "why is this still being done manually?" For software teams, that created a wave of practical business apps: less glamorous than consumer apps, but often more valuable. The best products were the ones that removed friction from daily work. See enterprise app patterns after remote work.
-
Apple Silicon
The Mac became a development machine again. Apple's move from Intel to its own silicon was not just about battery life and faster laptops. It changed the way developers build, test, and think about performance across iPhone, iPad, and Mac. Native apps suddenly had a much clearer future, and cross-platform tools had to adapt to a world where mobile and desktop were no longer so far apart. For developers, it meant faster builds, better simulators, and a reminder that hardware still matters. Apple announced the Mac transition to Apple silicon at WWDC 2020.
-
Privacy Labels
Apps had to start explaining themselves. With privacy labels, the App Store began asking developers to disclose what data their apps collect and how that data is used. It was not perfect, and it still relied heavily on developer honesty, but it changed the conversation. Privacy stopped being a hidden technical detail and became part of the product page. A good app now needs a clean interface, a useful feature set, and a data story that users can understand before they download.
-
Small Business App Store Program
For smaller developers, the App Store became a little more survivable. Apple reduced its commission to 15 percent for developers earning up to $1 million per year. That did not end the debate about app store fees, but it mattered for independent studios, small SaaS products, paid apps, and niche subscription tools. When margins are thin, the difference between 30 percent and 15 percent can decide whether a small app keeps improving or quietly disappears. Read our take on Apple's small-business fee cut.
-
App Clips: Instant Access Without the App Store Download
Do not underestimate iOS14's App Clips. They are small apps (less than 10 mb in size), that don't need to be downloaded from the App Store. They load instantly, and are catered to one task, on premise. They are launched by either bluetooth, a QR code or an NFC Tag. They are not full featured apps, however they're a small part of an app with one focus centered feature. No need to register an account if the feature requires it, because all you need is Apple Sign-in and Apple Pay to make a transaction on the fly. Reasonably, we at underlabs do develop App Clip Apps. See our full App Clips guide.
-
Mixed Reality: Where AR and VR Meet Practical Training
What happens when you mix Augmented Reality (AR) with Virtual Reality (VR)? Magic without the motion sickness. AR has prevailed upon a lot of development these past years (which we have written about) and the credit for this exclusively goes to its integration in mobile applications. Large corporations and government institutions are scrambling to invest in their own AR application, such as the U.S. Government and Walmart. Walmart as an example has been using mixed reality to train employees and staff members. The results are not out yet, though rest assured we will surely comment on it as more information gets released.
-
Low-Code Development: Useful, but Not a Replacement for Custom Apps
Building an app via building blocks can only get you so much. They are great for converting an excel sheet into an on-the-road mobile app to contribute and consume data, however it has its limitations. As mobile app development features and demands get more and more demanding with complications, there is a lot of room for low complexity and straightforward features to become a common place to build via drag and drop blocks to release a simple solution for tracking inventory, and converting basic forms into an app. We will see more of these services come on line, and they certainly have their place, but they will not replace enterprise app solutions.
-
CubeSats: Small Satellites, Big Data Opportunities
These are tiny satellites (10 cm x 10 cm x 10 cm) that are relatively inexpensive to build. They have become more feasible for academic researchers and private individuals to launch such light weight (1.3 kg or less per cube) into orbit by the myriads or launchers (thanks to SpaceX). They can be used for real-time imagery, communication satellites, and other experiments that can be beneficial from space.
-
Quantum Computing: What Happens When Compute Power Changes Everything?
While at least a decade away, they differ from current binary computer is their Qubits; These are fundamental to quantum computing and are somewhat analogous to bits in a classical computer. Qubits can be in a 1 or 0 quantum state. But they can also be in a superposition of the 1 and 0 states. They can technically disrupt non-quantum resistant blockchains. However they have immense advantages to computational intensive processes, such as in biology for DNA sequencing, or Machine Learning, Financial Portfolio Optimization, etc;.
-
Blockchain: Databases Without a Single Point of Trust
Databases have a trust issue. The system administrator or developers working on the project can technically modify them. Databases hosted by centralized servers can be modified by their owners. Imagine a database (blockchain) where no one individual, company or government can modify. A distributed network that constantly verifies cryptographically all historical records and ongoing transactions is what Bitcoin's Satoshi Nakomoto promises. A cryptocurrency is one use-case, however it can be applied to other fields, such as supply chains, digital identities, food safety, voting, healthcare, notary registrations, and a lot more;
-
Autonomous Drones: New Use Cases Beyond Flight
Quadcopters have become phenomenally trendy, notably due to the realization of their potential uses, especially when autonomous (unmanned). Underlabs has recently commenced flying and programming sequence tests for the sake of playing around. As Transport Canada finalizes the full requirements and safety regulations, Canada has the potential to innovative in such types of UAVs, in a myriads of fields. Think of expedited maintenance, surveying construction sites (3D models), light payload deliveries, and others more.
-
AI and Machine Learning: From Prediction to Decision-Making
Artificial Intelligence (AI) is when machines mimic human cognitive functions such as learning and solving problems. Such capabilities are penetrating many traditional industries and innovative solutions spring up, as we see today in health care, the automotive industry, language translations, chat-bots & intelligent assistants, non-linear controls, robotics and many more. As over the counter computers reach quantum computing, answers to questions would be instantaneous. Would we be happy with the answers?
Thats up to the past to tell. -
Self-Driving Cars: The Software Behind Autonomous Mobility
Self-driving cars (autonomous cars) are able to be "aware" (loaded word) of their surrounding due to camera sensors (computer vision), lidar, GPS and odometer in order to navigate paths and identify obstacles and road signage. The promise is the reduction in human car accidents, increase in road capacity, important to every major city mayor, solving the lack of parking spaces in cities. This could enhance traffic, and reduce the cost of transportation. Many technological challenges and lack of legal frameworks still exist, however the potential is very much real.
-
Apple unveils Watch
Samsung has the Gear 2 & Apple unveilled Watch. These wearables such as Apple Watch will not only give us the time, but will also help us make the most of it. With Apple's recent WatchKit, developers will race into this new platform to integrate a whole leap of features which would enhance and work in conjunction with their iPhone Apps; And Underlabs is at the fore front of this race.
-
Pay On Location & In Apps
Paypal isn't happy. A new competitor to the payment processing system comes from a staunch company, Apple. This may begin a price war with other Payment processing systems. Apple Pay may even be integrated into Apps to process transactions (not to confuse with In App Purchases). However for large scale adoption, merchants have the upper hand in integrating Near Field Communication into their Point of Sale. Mobile Payments is expected to reach US$70-billion by the end of this year.
-
Material Design: Motion & Interaction
Google has just unveilled today Android L, with Material Design. A guide for visual, motion, and interaction. The new theme puts iOS7 & Android fairly on a flat surface. While the trend continues to be fat & flat, material design offers a skeuomorphic motion to user interactivity. Expect slick designs and subtle effects coming out of Android apps.
-
The Retailers dream: Beacons
Imagine walking into a clothing store, your phone notifies you of whats on discount, save your favorite items, and head out for later viewing. No annoying sellers pushing fashion on you. Beacon is a precise in-door GPS, a micro-location targeting system. Real world scenarios such as Museums, Stadiums, Retail stores, etc; can notify their guests about any content information in a beautiful, rich way. As your customers browse through a store or mall, the items they're looking at would virtually appear on their screens; Allowing them to favorite, order, or make a purchase.
-
The Internet of "Everything"
Your home lighting, your TV, your thermostat, your car, many technical field equipments, are all going to be connected to the internet and run on software. They know how to talk to each other, which makes them smart. The intelligence is in utilizing them for the benefit of society and simplifying daily routines. At underlabs, we have built prototypes in regulating Hue Lamps automagically when Montreal's gloomy days hit. It turns our cloudy days into sunny scenes.
-
Hybrid Cloud & IT as Service Broker
Enterprises are beginning to design private cloud services with a hybrid future in mind and make sure future integration/interoperability is possible.
The need, and necessity to expand into hybrid cloud computing from private cloud services are taking on the CSB role. Terms like "overdrafting" and "cloudbursting" are often used to describe what hybrid cloud computing will make possible.
-
Cloud, mobile, social & information
Four major forces are revolutionizing the way applications are developed and used — cloud, mobile, social and information. They’re overlapping and amplifying one another to dramatic effect. Suddenly it’s a mobile, social, cloud-based, real-time world. Some enterprises will recognize the moment and seize the opportunity to gain a massive competitive advantage over those that do not. The individual pieces are important, but you’ve got to understand how they all fit together as well.
-
Development Iteration & Lifecycle
Five years ago, you might plan on a year to build an app. Today, if you can’t turn around a new application in a month, you’re not in the game. If you’re not using agile and cloud, you have no way to build out apps quickly enough to differentiate yourself from your competitors. Cloud lets you scale up — and shrink back — quickly. Even the funding model has changed. Capital investment is too slow; cloud services are operating expenses.
-
B2C Interactions
The goal is to interact with a customer in real time, then there will be the seamless need for integration across Web, mobile devices, cloud and social media. All these different user experiences, applications, platforms, processes and access points must be tied together. No matter how advanced we get in the various disciplines, ultimately it has to come together in one view, integrated under one cohesive strategy. Underlabs covers the breadth of application disciplines, tools and strategies on the basis of trusted independent research from Gartner.
-
Enterprise Rich Internet Application
RIAs are the ideal tool for combining and presenting the objectives of a group of distributed business actors who need to exchange accurate and timely information. The rich functionality and interactivity of modern internet applications, combined with underlabs's experience in the development of enterprise applications creates powerful business solutions.
-
Data Visualization: Turning Business Data Into Decisions
Through intelligent and creative design and the use of RIAs, underlabs can present your valuable data in a format that makes sense - providing rich and interactive insight into that data for dispersed audiences. By combining our team's development talent and creative flair, underlabs builds dashboards and data views that bring your data and business trends to life. It's your data, get the most out of it.
-
The Mobile-First Company: Why Every Business Needs a Mobile Strategy
The App Market is a New World, to everybody. There are emerging ways to tap into an exponentially increasing market which gives access to everybody's pockets and eye attention. They provide the medium to transform everything we know about traditional media, entertainment & service.
-
Isometric 3D Virtual Worlds
From the hobbyist to the veteran game developer, if you want to make an isometric game in Flash; Possibilities for virtual goods, MMO games, and story lines can go on and on... in your own little world.
-
Web Apps with Google Toolkit
With cutting-edge development tools, frameworks and libraries such as GWT Google Web Toolkit and advanced development methodologies such as agile and extreme programming assists speed up development and shorten time-to-market cycles.
-
GIS Solutions: Making Location Data More Accessible
Today we are able to access rich geographical content from maps, survey information and CAD drawings through to revolutionary Immersive 'Street View' videos. Traditional GIS solutions can be large and expensive, bound by their Client-Server architectures, whereas underlabs browser-based GIS viewers make your data accessible where and when you need it.
Go deeper
Curated reads from Apposphere when a timeline entry needs more context.
Planning an app
I Have an App Idea. What Should I Do Next?
AI in real products
Profits in AI is Now at the Application Layer
Automated Workflow Architecture with LLMs
Launch readiness
Vibe-Coded App Rejected?
Platform shifts
App Clips guide
Spatial Computing / Vision Pro