How Siri Got Off Track – The Information
The Information’s Aaron Tilley and Kevin McLaughlin have a 4,000-word article (N.B. paywall) detailing the checkered history and infighting that has characterized Siri since its launch in 2011 on the iPhone 4s. The article starts with some background:
“Many of the former employees acknowledged for the first time that Apple rushed Siri into the iPhone 4s before the technology was fully baked, setting up an internal debate that has raged since Siri’s inception over whether to continue patching up a flawed build or to rip it up and start from scratch. And that debate was just one of many, as Siri’s various teams morphed into an unwieldy apparatus that engaged in petty turf battles and heated arguments over what an ideal version of Siri should be—a quick and accurate information fetcher or a conversant and intuitive assistant capable of complex tasks.”
Siri the Information Fetcher
For anyone that has used Siri, the “information fetcher” group apparently won out. Siri has never risen to the level of managing complex tasks and is the key reason why so many people compare it unfavorably to Amazon Alexa and Google Assistant. In its defense, Siri is now quite good at executing a narrow set of tasks on the iPhone. However, that was not always the case. When it launched, there were many complaints about Siri’s ability to properly interpret accents, perform any tasks or meet expectations. In Episode 33 of the Voicebot Podcast, Mark Webster, CEO of Sayspring, commented (35:16) on Apple’s mishandling of consumer expectation settings in 2011 and contrasted that with Amazon’s Alexa launch in 2014.
What Alexa did really well when came out…[is] we were told we could do three or four things and she performed those really, really well, and that was a great experience. Siri came out in 2011; much earlier from a tech perspective. But, we were told you could do anything with Siri and that wasn’t the case and it was super disappointing.
Richard Williamson, an Apple executive that became responsible for Siri shortly after it’s launch, seems to agree with Webster’s assessment of Siri’s initial introduction. The Information article quotes him as saying:
After launch, Siri was a disaster. It was slow, when it worked at all. The software was riddled with serious bugs.”
The Need to Open Siri Up to Developers
Voicebot has long suggested that Siri must open up to developers in order to move forward. Apple’s tight control of hardware and software has been critical to implementing great user experiences on a variety of devices. However, these devices were all programmatically limited. There was a set number of features that could be performed and the boundaries were clearly delineated. Artificial intelligence (AI) changes the boundary set for what devices can and are expected to do. Yes, Apple uses AI for tightly defined services such as facial recognition and music discovery. However, Siri falls short precisely because it can only perform a limited number of functions in nine specified domains. AI-driven voice assistants should do much more.
“The absence of [strong] leadership and the constant turnover has held Siri back in key ways, these former employees said, most notably in the failure to open up Apple’s notoriously closed culture to allow outside developers a greater opportunity to create a broader array of useful Siri apps…Apple in 2016 introduced SiriKit, a tool that allows outside developers the opportunity to create new capabilities for Siri. However, it hasn’t yet led to the explosion of developer interest that its founders envisioned.
“The original vision was for Siri to be an intelligent digital assistant capable of responding to a user’s request and also interacting to get the desired response…Siri’s founders believed that the only way to achieve this result was to create an open platform that allowed outside developers to integrate into Siri. By doing that, the assistant’s capabilities could scale far beyond what the team could build in-house.”
That strategy sounds exactly like what Amazon implemented with its Alexa developer community and that Google is furiously trying to replicate for Assistant. The article anonymously quotes one of Siri’s original founders as saying:
It was about making the App Store for AI. It was supposed to be a way to orchestrate the internet through conversation.
Technical Debt and the Rise of Competitors
Siri’s capability set seems to be arbitrarily constrained unlike Google Assistant and Alexa. However, former Apple executives suggested that there was considerable risk in opening up Siri to developers because the “software was so brittle and inflexible.” SiriKit was started in 2012, but wasn’t released to developers until 2016.
Google and Amazon invest heavily in a broad range of voice assistant features and count thousands of independent developers as partners building out ecosystem capabilities. The manpower multiplier of partners is valuable on its own. However, that is only part of the equation. More voice apps and more users result in more interaction data around more use cases. Those interactions all become training data that can improve the AI behind the voice assistant as well as AI services for anticipating needs and guiding users to better overall experiences. The original technical constraints of Siri may be creating a devastating long-term disadvantage. This is an example of how technical debt can become so large that starting over is the only option.
A Revolving Door of Siri Executives
If there is a unifying theme behind Siri’s struggles in The Information article, it is that a lack of consistent, visionary and authoritative leadership is to blame. Executive leadership on the Siri team is characterized as a revolving door with the exception of Bill Stasior, VP of Siri and Search, and Alexa Acero, Senior Director of Speech, who joined the team in 2012 and 2013 respectively. Another undertone is the second-class status of services compared to hardware at Apple. That sentiment was expressed by former Apple employees as a reason why Siri’s challenges were not seen as important enough to clean up.
Think about this. When Siri launched, its team was just 100 people. Amazon has over 5,000 employees dedicated to Alexa with nearly 1,500 additional job openings today. It is not clear how many people are dedicated to Siri, but given its current state, the number is not nearly enough.
HomePod Reviews Prompted the Story
Finally, it is worth noting that The Information set out to investigate the Siri story after seeing so many poor reviews of the voice assistant related to the recent HomePod smart speaker launch. The reporting concludes:
“Siri is a problem. It’s arguably the main reason that Apple’s latest product launch—a $349 smart speaker called HomePod—has underperformed, based on early estimates from analysts. Although the speaker won plaudits for its sleek look and audio quality, review after review trashed the Siri functionality.”
Voice is the New UI – Expectations Have Changed
No matter how great the hardware may be, the user interface (UI) for a smart speaker is the voice assistant. Apple may be known for hardware, but it is even more revered for its dedication to simple, reliable and friendly UI innovation. The MacIntosh introduced the mouse-driven point-and-click UI. iPod gave us the click wheel for navigation. iPhone may not have introduced the first touch-based mobile interface, but it was the first that consumers loved. This is about more than smart speakers. Siri will be the primary UI for Watch and AirPods too. Maybe if Apple recategorized Siri as a UI instead of an assistant service, it would finally get the attention it needs.
Apple has many advantages still at its disposal if it can put Siri on the right track. Siri is accessible on more than 500 million devices, claims 375 million monthly active users, supports 21 languages and processes “two billion requests per week.” All of the these numbers exceed official usage announcements for both Google Assistant and Amazon Alexa. The question is breadth of use and long-term prospects. Expectations have changed and Siri must also.