We set out to build a new kind of search engine. It’s only recently been launched but it’s actually working as intended—better, in fact. Our engine returns answers. In fact, Jelly returns a direct answer to 95% of questions with a median time of 14 minutes. It really works!
How Jelly Works
There’s an old saying, “We all die with our stories still inside us.” Jelly uses computer science to unlock the unimaginably immense amount of data in the form of knowledge, opinion, and expertise that lies within most of us. This data is not published online — but it’s just waiting to be released.
Questions come into Jelly anonymously just like classic Web search. But, that’s where all similarity ends. Jelly questions go through a process in service of returning a quick, simple, and reliable answer from a person who has volunteered to respond in areas of topical interest and experience.
Building on Something of Value
The system we put together is just a start, but it’s something solid that we are already improving. For example, next week we will be submitting a new build that leverages what we’re calling, Instant Answers to substantially lower that median time and improve overall experience.
Instant Answers are high quality answers delivered in a couple seconds. Tests indicate about 18% of questions can be assigned a high quality Instant Answer. If an Instant Answer misses, Jelly does what it was built to do, route that query to the right person for a custom written, on demand answer.
Jelly gets smarter and faster with every question submitted. One of the most interesting attributes of Jelly is its ability to learn from its own mistakes. If an Instant Answer does miss, as I said, Jelly does its thing, it provides an on demand answer. This new answer helps Jelly learn and makes it better.
Timing Is Everything
With the ascendence of Conversational UI, that is, user interfaces based on speaking aloud to a device, Jelly becomes invaluable. The idea of pushing someone a long list of hyperlinks to documents is absurd in this new scenario. What’s needed is the answer—simply and directly.
All of these conversational UI products coming out match my personal vision for the future of technology. Personally, I think technology should be mostly out of view but always there when you need it. I’m also of a mind that technology should amplify the best traits of humanity.
After building and launching Twitter, I declared, “If Twitter succeeds, then it will be a triumph of humanity, not a triumph of technology.” Recently I said, “The world needs Twitter.” We’re heading into new territory and I’ve now come to realize, the world also needs another way to get answers.
Now that we know Jelly can work, and what it may work especially well for, we certainly have our work cut out for us. Ben and I believe that a small team can have a huge impact. But, not as small as we are now. We’ll need more than 5 engineers where we’re headed. Want to join us?
Co-founder and CEO
Jelly Industries, Inc.