Apple to acquire voice app firm PullString in deal worth below $100M, report says
As part of a larger push into voice recognition and artificial intelligence technology, Apple has reportedly agreed to purchase PullString, a San Francisco startup focused on the development and publication of voice apps for Amazon's Alexa, Google Assistant and "internet of things" hardware.
PullString Converse's conversational "AI canvas."
Citing sources familiar with the matter, Axios reports Apple's deal with PullString is worth under $100 million, though executives have the potential to gain larger payouts. As with most Apple acquisitions, the PullString deal is believed to target talent and technology, which in this case involves voice apps.
Formerly known as ToyTalk, PullString was founded by former Pixar executives in 2011 when voice-powered applications were still a novelty. Specifically, the firm's main product, Converse, is a voice design and artificial intelligence tool that allows customers to create "expressive apps" with custom voices and advanced sound design, according to its website.
Converse saw initial integration in toy designs from Mattel, including interactive Hello Barbie and Thomas the Tank Engine products. More recently, however, PullString moved into IoT devices with a focus on virtual assistants like Alexa and Google Assistant. The current generation of PullString Converse enables users to build out customizable, conversational interactions with users, a feature that Apple's own Siri product lacks.
A list of successful Converse integrations on PullString's website includes integrations like trivia and quiz experiences, choose your own adventure games, voice-enabled FAQs and customer support experiences aimed at troubleshooting product issues before escalation to a human operator.
What Apple has planned for PullString is unknown, though the firm's past endeavors strongly suggest Converse technology will be applied to bolster Siri's feature set. Whether that strategy involves a more developer accessible Siri is unknown, but the technology behind Converse would imply support for outside parties is a distinct possibility.
Despite being first to market with a voice assistant, Apple's Siri solution has in many ways been surpassed by open platforms like Alexa and Google Assistant. An abundance of voice apps -- called Alexa Skills by Amazon and Google Actions by Google -- powered by AI backends presents users of competing platforms access to a vast array of functions and services. Siri, on the other hand, is viewed by many as wanting.
Converse could help change existing perceptions by presenting developers the requisite tools to build advanced Siri-compatible apps. Alternatively, Apple could use PullString's toolset to help its own Siri software engineering team bake unique first-party functions into hardware interfaces like iPhone and HomePod.
PullString Converse's conversational "AI canvas."
Citing sources familiar with the matter, Axios reports Apple's deal with PullString is worth under $100 million, though executives have the potential to gain larger payouts. As with most Apple acquisitions, the PullString deal is believed to target talent and technology, which in this case involves voice apps.
Formerly known as ToyTalk, PullString was founded by former Pixar executives in 2011 when voice-powered applications were still a novelty. Specifically, the firm's main product, Converse, is a voice design and artificial intelligence tool that allows customers to create "expressive apps" with custom voices and advanced sound design, according to its website.
Converse saw initial integration in toy designs from Mattel, including interactive Hello Barbie and Thomas the Tank Engine products. More recently, however, PullString moved into IoT devices with a focus on virtual assistants like Alexa and Google Assistant. The current generation of PullString Converse enables users to build out customizable, conversational interactions with users, a feature that Apple's own Siri product lacks.
A list of successful Converse integrations on PullString's website includes integrations like trivia and quiz experiences, choose your own adventure games, voice-enabled FAQs and customer support experiences aimed at troubleshooting product issues before escalation to a human operator.
What Apple has planned for PullString is unknown, though the firm's past endeavors strongly suggest Converse technology will be applied to bolster Siri's feature set. Whether that strategy involves a more developer accessible Siri is unknown, but the technology behind Converse would imply support for outside parties is a distinct possibility.
Despite being first to market with a voice assistant, Apple's Siri solution has in many ways been surpassed by open platforms like Alexa and Google Assistant. An abundance of voice apps -- called Alexa Skills by Amazon and Google Actions by Google -- powered by AI backends presents users of competing platforms access to a vast array of functions and services. Siri, on the other hand, is viewed by many as wanting.
Converse could help change existing perceptions by presenting developers the requisite tools to build advanced Siri-compatible apps. Alternatively, Apple could use PullString's toolset to help its own Siri software engineering team bake unique first-party functions into hardware interfaces like iPhone and HomePod.
Comments
Good buy!
But when will Apple use their tech? Apple acquired a company years ago that had a voice assistant that was so smart it makes all the current assistants look like 4-year-olds.
As for your question, I’m thinking that they’re going to need some really good voice AI by the time their AR glasses come out. With pinpoint binocular eye tracking, voice will be the way we “click” on an object in the depth field to select it for more information or whatever. Dialog with the data base on the object of regard will have to be ready for this platform. Revolutionary, needless to say.
YES. VocalIQ.
https://www.businessinsider.com/how-apples-vocaliq-ai-works-2016-5
"For example, imagine asking a computer to "Find a nearby Chinese restaurant with open parking and Wi-Fi that's kid-friendly." That'd trip up most assistants, but VocalIQ could handle it. The result? VocalIQ's success rate was over 90%, while Google Now, Siri, and Cortana were only successful about 20% of the time, according to one source."
...Acquired over 3 years ago.....
I'm just wondering what the heck they're doing? I can only imagine they're building a gigantic Siri rollout.
I do remember reading about it, and it sounded great. Who knows, maybe they will surprise us with Siri 2.
if I ask Siri, “How old is Ben Kingsley?”, get the answer then ask, “Who is his wife?”, get the answer then ask “What is her bra size?” Siri comes back with “Here’s what I found on the web for ‘bra size’.” Complete fail! That proves Siri sucks compared to all others even though I haven’t tried this exact same sequence anywhere else. I heard Siri sucks so it’s true because of internet.
/s
I’m with you. I use Siri multiple times a day for a variety of tasks and rarely have issues. When I run into problems it’s usually because I have a poor connection.
Exactly. Even though she is limited on the HomePod, I find her especially rewarding because she hears great. She does all the things you mention and more. She increased he joke library to boot. She added Wikipedia to her knowledge base. I ask her things all the time like, "who whom the Oscar for best picture in 1972." She knows the answer. She also can tell you things like what is the best reviewed coffee place or what is the closet coffee place. She is completely useful. She keeps improving without announcements to For instance, Siri on HomePod couldn't originally do multiple timers. Now she can. I often tell her I love a song, which means she will remember what songs are my favorites. Siri will make reservations.
She sometimes will fail on my phone if my phone is on wifi and I have a poor connection or if I am using her over bluetooth in my car but this is for poor internet connection issues. Some of this could be improved by not needing the internet for certain tasks that could be performed on the phone.
Siri also needs works building off previous asked questions.
Apple also hired a person to lead their smart home efforts:
https://www.cnbc.com/amp/2019/02/16/sam-jadallah-to-join-apple-as-its-new-head-of-home.html
Amazon has an army of 10,000 people working on Alexa. Aquiring Pullstring has just put Apple into a position to level the playing field and provides a platform build a SiriOS. So, much for ‘Apple will never catch up’ ¯\_(ツ)_/¯
This should give Apple a good head start in developing a great tool for their SIRI partners.
The tool could be packaged separately or integrated into Xcode. Time will tell.
Here is a nice example in the health care field which Apple is so interested in.
Here is an interesting presentation at Google I/O 2017
Also I can see how this team was a better fit for Apple than Google. More focus on philosophy & arts than technical & engineering.
You mean that siri works decent if you use the exact examples Apple says to use in the exact right phrasing?? Shocking. And quite a low standard for something Apple should be years ahead on, not what seems to be years behind.
But for me, it doesn't even do those basic things right. For example, I have to edit the majority of reminders I set. And have you ever tried to dictate more than a 3 word message to someone and not run into problems? You're kidding yourselves. There are so many examples of every day things siri gets wrong.
Siri is incredibly dumb. No just not sophisticated, but just plain dumb. To pick one example, if you ask siri to define a word, she'll define it (easy, rote task), and then she'll ask, "do you want to hear other definitions?" If you say "no" because you want to read and review the definition on the siri screen, siri will push the definition off screen and show only the word "no".