The iPhone was the first smartphone to include a digital assistant, but within a few years, competitors have completely surpassed Siri's capabilities. But Apple (AAPL 1.85%) continues to include Siri in more and more products -- most recently the Apple TV.
Meanwhile, Google (NASDAQ: GOOG) (NASDAQ: GOOGL) developed Google Now, which predicts what information a user wants and pushes it to their notifications, or displays it under its search bar. Microsoft's (NASDAQ: MSFT) Cortana has proven its worth, and it's built into Windows 10. Facebook (NASDAQ: FB) has developed a personal assistant for Messenger that's more of a concierge service than an appointment scheduler.
As digital assistants become more prevalent and are expected to do more for their human overlords than ever before, Apple is facing pressure to keep up with the competition. A recent Reuters report found that Apple is currently looking to fill a whopping 86 positions related to artificial intelligence. Reuters reported that Apple doesn't say how many people it has working on machine learning, but said one former Apple employee estimated the number to have tripled or quadrupled in the past few years.
The Apple challenge
When it comes to understanding humans, Apple's AI team is presented with a unique challenge. The company values its users' privacy much more than Google, Microsoft, or Facebook. As a result, Apple only collects a limited amount of data on its users.
That makes determining what people want somewhat difficult. If you can't track how people respond to Siri, and use that data to tweak her responses, Siri quickly becomes less efficient at tasks compared to Google Now or Cortana.
Indeed, the entire basis of Facebook's digital assistant, M, is to test it with a small group of people, and use their responses to scale the service. It's the entire basis of deep learning AI -- use big data sets to provide better results.
But Apple won't go for it. This presents a challenge for Apple and its software engineers. In fact, some engineers may shy away from Apple due to the heavy constraints on user data. With plenty of other big tech companies entering machine learning, there are plenty of other jobs with better access to data to work on.
But Apple is a device company
All three of the competing tech companies mentioned in this article are service companies supported by advertising. It makes sense for them to spend heavily to develop another service that could benefit their advertising business through increased search volume or connecting businesses to customers. Apple, however, is a device company -- it makes money when selling a device to a consumer.
But the key to Apple's products is that it uses software to differentiate its devices. Other companies (like Xiaomi) may be able to copy its hardware designs, but they can't touch its software.
However, Google Now, Cortana, and M are all available on both iOS and Android, with the latter two available on Windows Phone as well. Now that competitors have superior capabilities, Siri is no longer a differentiator -- it's an also-ran feature that many users will opt to replace. What's more, the experience of using Google Now on Android or Cortana on Windows Phone is vastly superior to using either on iOS.
And that's a serious issue for Apple. The iPhone represents over two-thirds of Apple's sales in fiscal 2015. As sales come up against tougher comparables (following the one-year anniversary of the hugely popular iPhone 6) later this month, Apple needs to maintain all of its advantages in smartphones to continue growing its sales.
But Apple has a plan if it can't make Siri as good or better than the competition. Some of those machine-learning experts are going to work for its product marketing and retail divisions. Maybe they can optimize retail sales while Apple works on Siri.