Stumped Siri? Soon She Could Crowdsource Tough Queries

Apple has filed a patent that explores a crowdsourcing solution for when Siri can't answer your questions.

Siri is not known for her people skills.

For years, iPhone owners have been documenting funny or frustrating examples of times when the Apple personal assistant failed to come up with a suitable answer to a question. But what if, when Siri failed, she could call upon the crowd to find the right answer?

That's the premise of a patent Apple filed with the U.S. Patent And Trademark Office, as spotted by AppleInsider. The patent, titled "Crowd sourcing information to fulfill user requests," documents a system through which Siri, upon failing to deliver a satisfactory answer to a user's query, could crowdsource the desired response to "address questions previously unanswerable in real time." According to the patent, Siri would crowdsource information from various sources, including general websites, Q&A forums, and, most interestingly, real people. Using human sources could give Siri access to real-world experts who can easily understand and answer questions Siri can't.

Interestingly enough, another company has already explored the concept of a crowdsourced Siri: Sivi, a now-defunct paid service we wrote about last year that employed human concierges who scoured the Internet to answer users' incoming questions.

[Image: Flickr user Robert S. Donovan]

Add New Comment


  • David McSwain

    Startup, aims to do much the same except focused on teams and companies. It allows teams to build their own voice-enabled question answering sites to link spoken questions or phrases to wiki pages, photos, videos, Sharepoint or Google Docs. Any URL internet or intranet. It supports questions like "What were widget sales for last quarter?" or "Retrieve expense form". The answers to these are very team/company specific and thus require a crowdsourced approach.

  • Suleman Ali

    Perhaps there should be a list of the standard difficult questions which Siri is asked. That may be interesting in itself in discovering what computers do and don't know.