A Look at the iPhone 16 with Onboard LLM, AppleGPT

Training on-device LLMs with private data is the future of personalized brand experiences

Image and article copy by Rehab AI Consultancy

tl;dr: Apple's latest iPhone feature, the onboard AI 'Ajax' is about to redefine personal tech. Think AI that not only understands you but also keeps your data private, directly on your device. How will this transform your daily interactions and experiences? Imagine the possibilities – they're closer than you think.

An introduction to AI agents and on-device LLMs

As predicted by our AI consultancy team (2023 Whitepaper), the next era of AI is AI agents running on devices. To break that down, agents—instead of asking AI to perform a specific task directly—are trained to figure out the right tasks to achieve a specific result or objective. On-device, instead of accessing AI in the cloud, smaller models with less energy requirements run on your local device, specifically your smartphone.

Enter the elephant in the room: Apple. It’s the only company not to launch an AI product or service in this LLM era—that is, until this year. Apple has been developing its own AI LLM called Ajax, which is similar to ChatGPT and also has multimodal and image generation capabilities.

It's because of this we think the upcoming iPhone 16, iOS 18, Apple GPT (Ajax), and Google Gemini will be the AI story of the year. Let me explain.

Current LLM capabilities are all around the Open AI GPT4 level. Yes, there are slight differences between Gemini, Claude, and OpenAI, but they are much the same. They start to come to life when you train, fine-tune, or give them access to your data. What's different about what Apple can now do is that it has the ability to run on a phone without needing access to the internet. When you run on a device the LLM is on the phone (Called Edge AI), there's no need to fear that you're uploading data to the cloud and training the model with your personal or commercial data. Plus, the responses are more real-time, providing information to you not as a response to a prompt but proactivly based on the data on the phone.

This is important because it starts to solve one of the biggest fears around adoption, which is data privacy, which is a transformative moment that will reshape customer service, marketing strategies, and consumer interaction forever. 

The core concept: onboard AI and user-centric learning

Central to the new iPhone is its onboard AI, meaning it's on the phone. You could be disconnected from Wi-Fi and cellular data and the model would still work. Additionally, your smartphone has a treasure trove of data and information that is currently unavailable to online models like ChatGPT.

We believe that this personal data, from your location, calendar, messages, and behavior, is the true way to create intelligent, personalized assistants, something the industry has been trying to deliver for the past ten years (Siri, Alexa, Google Assistant).

Choose which data you want the onboard AI to access

Why this is such a big story is because this technology is being provided by Apple, which has such a strong ethos around privacy and security. Apple is probably the brand that most consumers would allow to access its data on-device.

We believe that, much like existing permission settings for microphones or cameras, this AI will have similar settings, allowing users to control their access to various data sources on the phone. This approach ensures a balance between personalization and privacy, enabling the AI to tailor its functionality to the individual's preferences and behaviors. Enter brand utility, the ultimate brand engagement tool.

The onboard AI is poised to unlock a hyper-personalized data layer for brands. It provides unprecedented insights into consumer likes, wants, and behaviors. By granting iOS applications access to the onboard Large Language Model (LLM), brands can leverage this data to enhance user experiences, offering more relevant information and customized interactions.

These intelligent, AI-driven tools have the potential to create a deeper, more personal bond between brands and consumers. Brands can transition from being mere providers of goods or services to becoming integral parts of their customers' lives. This closeness brings a new dimension to brand loyalty, moving beyond transactional relationships to truly personal ones.

Case study: Nike Running Club

Let's look at the transformation of an app like Nike Running Club. With access to the onboard LLM, a Nike Running AI Agent could offer highly personalized running recommendations via Spotify, product suggestions, and fitness advice, all tailored to the user's health data, schedule, music preferences, and habitual routes—all based on the user's goal. Be that to stay healthy or train for a marathon.

The Nike Running AI Agent, a personalized proactive experience

  • Health Data-Driven Training Routines: The onboard AI's ability to access and analyze health data from the iPhone will create an entirely new type of personalized training. By understanding a user's fitness level, heart rate trends, and other data like VOMax and recovery time, the AI Agent will be able to create tailored workout plans. E.g, if the AI detects that a user's heart rate is consistently high during runs, it will create new routines that focus on building endurance. Or if you are recovering from an injury, the plan would be adjusted for low-impact, accelerating the return to fitness.

  • Location Data for Personalized Running Routes: The utilization of location data opens up exciting possibilities for customized running experiences. For female urban runners, the AI agent could identify more crowded streets or parks, providing a layer of security and enhancing enjoyment. For trail runners, the AI agent would recommend new unexplored trails, adding variety to the training.

  • Product Recommendations: Based on the analysis of the user's running data, the AI agent can also provide assistance in ordering personalized gear. Shoes best suited for the user’s running style or apparel for the current weather conditions could be suggested, simplifying the shopping experience and increasing loyalty.

  • Adaptive Music Playlists, return of the Power Song: Finally, by allowing the AI to access the Spotify app via permissions, the AI Agent could craft the perfect playlist and include AI-selected Power Songs (this was a cool feature that no longer exists) for the run and play them when the user starts to lose energy.

Allowing the onboard AI to train itself on the user's health and location data delivers an experience never seen before because we have never been able to train AI on a specific user's data. Delivering a highly individualized and immersive running experience.

Revolutionizing customer service

The last decade witnessed a decline in customer service quality, primarily due to cost-cutting and offshoring of support services. The onboard AI in Apple's new iPhone could revolutionize this landscape. 

Now customer service experiences can be personalized, real-time and proactive, the AI agent knows the consumer as well as the personal shoppers in luxury retail brands. This technology could enable brands to scale personalized service, transforming customer interactions across all touchpoints.

While all consumer brands stand to gain, commoditized sectors like telecoms, banking, finance, and insurance could see the most significant impact. These areas, often criticized for mediocre experiences, could benefit from enhanced personalization and proactive customer service, which involves making changes and suggestions for the user rather than having to ask for them.

Finally, where is Gemini in all of this?

The rumor that Google Gemini will be integrated into the iOS platform adds a new dimension to AI's capabilities. We predict that Google will provide the real-time knowledge graph provided by the new and advanced models of Gemini. Whereas the Apple LLM will run on a device, Google Gemini’s LLM will be utilized to bring in data and information from the wider internet. The deal is probably akin to the $25+ billion dollar existing deal to make Google Search the default search engine on Safari. 


In this next chapter, your smartphone becomes a gateway to a world where brands understand you, help you and connect with you on a once unimaginable level. The future of brand interaction is here, and it’s personal, proactive, and profoundly impactful.

This blend of new technology and previously inaccessible data promises to redefine customer service and brand interaction, as well as set a new standard for user-centric technology. The upcoming iPhone's onboard AI could mark the beginning of a more intuitive, personalized, and proactive approach to technology.

Let us know what you think in the comments below, and find out more about Rehab AI Consultancy here