AI is maturing. But what does it mean for app developers now?

It’s been nearly two years since OpenAI launched ChatGPT and almost instantly made AI the topic of conversation in the C-suite, at the office coffee bar, and even around the dinner table.

Since then, AI and large-language models (LLMs) have been front and center in virtually every tech industry gathering. Business leaders have been closely tracking what the technology leaders — most notably Google, Microsoft, OpenAI, Facebook, and Apple — are saying and building with AI. 

After Google’s AI-laden keynote at Google I/O in 2023, we wrote about how as AI evolves, it will be built into more applications. We proposed that the initial ChatGPT model of a standalone AI-powered chat would fade and evolve into a background technology that is a part of and powers many applications. That certainly has proven true, as Apple unveiled its own version of “AI Inside,” dubbed Apple Intelligence, during this year’s WWDC. 

So, what’s different with AI this year? In a word, maturity. 

Another year of maturity means AI has become incrementally more useful in two different ways:

• As a tool for builders (like our product strategists, designers, and developers at ArcTouch) to help accelerate software development. 

• And as a market-ready technology component that can inject intelligence into our trusted applications and power new types of digital experiences.

Let’s take a deeper look at those two opportunities. 

1. AI as a technology component

Take Apple’s new Apple Intelligence, which will soon be available in some of the company’s newest iPhones and iPads, and eventually the Mac. Apple has a history of entering a market after other leaders — e.g. mp3 players, smartphones, wearables, etc. — and doing things a whole lot better than those who came before. Apple’s innovations often pave the way for mass market adoption and signify the maturation of a technology or product category.

In the case of AI, Apple’s approach is to build super-intelligent components into the company’s operating system, developer tools, and existing suite of applications on all its devices. “AI for the rest of us,” as the company boasts, promises to help people improve how they create, modify, and manage content. It’s an assistant — also built into Siri — to help write emails, edit photos, create personalized emojis, and virtually everything else we do on our Apple devices. 

Purpose-built apps with AI at the center

Apple’s interest in mature AI goes beyond making its own apps smarter. By announcing Apple Intelligence at WWDC, the company told thousands of app developers how they could leverage Apple’s new AI tools in their own apps. 

Google, of course, was talking about building AI applications last year at I/O 2023 — showing powerful integrations of generative AI in Gmail, Google Docs, and more. The company also offered an early API that gave developers the chance to build some prototype AI-powered applications. But it was so early. 

Another year offered maturity in the form of Gemini, Google’s rebranded and evolved suite of AI products, unveiled at I/O 2024. And a few weeks later, Apple Intelligence gave developers the confidence and the business case to start building AI-powered applications across platforms.    

So, here we are. As we head toward the end of 2024, developers have the tools they need to build AI-powered apps for iOS, Android, and beyond. Based on discussions with some ArcTouch clients, they will. We expect next year, most apps will have some kind of AI component built into them. 

It won’t always be noticeable. It might just be an AI enhancement to a text entry or one that intelligently surfaces content based on your behavior. These sometimes subtle improvements will make apps that you already depend on more intelligent and more lovable. We like to talk about apps as purpose-built — like steak knives designed to do one or a few things very well. AI has a chance to enhance these apps. To make them do that one thing even better. To sharpen that knife blade. 

Eventually, you’ll start to see some new types of experiences that simply couldn’t have existed without AI. But that’ll take a little more time — and maturity.    

AI that can power voice navigation and improve user engagement

As we wrote recently, the new and improved Apple Intelligence-powered Siri, along with the updated App Intents API, gives developers the ability to improve user engagement with their apps. 

Adding a voice interface to an app has long been an emerging user experience trend, dating back to the introduction of SiriKit, and the proliferation of voice assistants like Alexa, Siri, and Google Assistant. A recent voice search statistics report suggested that more than half of adults use voice search every day. 

More advanced AI and LLMs promise improvement in the ability to understand better and interpret our language and intents – and could be the key to unlocking full voice control of digital experiences. Today, we use Siri for quick searches and simple transactions. Tomorrow, we will use it to understand and handle complex multi-step requests. Voice assistants like Siri will orchestrate a sequence of interactions between several apps to achieve your desired goals. They will be conductors for app discovery and ongoing engagement

2. AI, the imperfect but helpful builder’s tool

If you follow our blog, you know that we’re bullish on the potential of AI and LLMs to make us better at our jobs. We have an ongoing series of articles in which we share some of our learnings as we’ve injected AI into our app development process. So far, we’ve shown how to leverage AI in our discovery and strategy process, in designing mood boards, and in writing user stories. 

So what does the maturity of AI mean for us as builders? 

For one, we’ve learned that the pace of innovation with LLMs is incredibly rapid. The results we get with our prompts improve dramatically with every passing week. In fact, we’ve struggled to publish these articles fast enough before some of the content is out of date. 

Fortunately, the main takeaways about using AI to build software still apply. As we wrote in our series intro:

• AI is a technology we can use to augment our work – saving us some time in certain areas so we can focus on others.

• That efficiency allows us to devote more of our time to tasks that deliver more value to ArcTouch and our clients — such as adding great new features, polishing our designs, and perfecting the user experience.

Most importantly, we’ve learned AI isn’t a turnkey solution for anything. As builders, we own the final software we release. It makes mistakes, and how it’s been trained or how and why it responds isn’t transparent.

We take responsibility and ownership for every aspect of those digital products. That includes work that any human on our team performs, any library that we license, or the output of any AI-augmented task.

A great illustration of the state of AI for builders comes from one of our favorite design tools. We wrote a blog post about some of the amazing new AI-powered features in Figma — only to see those features temporarily removed. The reason: A developer shared how he used Figma’s Make Design feature to create a weather app – and the result too closely resembled Apple’s Weather app. 

To reiterate: AI is a powerful tool to help augment how we build applications. But we have to use the same care and responsibility in applying its work that we would our own. Put another way, we can’t just steal someone else’s work just because AI serves it up to us.

AI and the buzz that won’t fade

Spoiler alert: This won’t be the last post we write about AI.

It’s changing how we work. And it’s opening up a world of possibilities in the products we build. 

It’s maturing. We’ll keep building with it, using it, learning about it, and, yes, writing about it.  

Our AI journey promises to be transformational and maybe a bit messy. And we’re excited for every bit of it.