There’s been a lot of news about ask.com today, with the announcement and launch of a new interface for the search engine. Some overviews of the new look, and reborn search engine:
- Ask Relaunches: Now “Ask 3D”
- Newly Relaunched Ask.com Glimpses of Greatness and Some Problems
Makes you wonder what else might have changed, or will change at ask. Here’s one future possibility:
Imagine searching for a laptop computer. You go to ask.com and type in a brand name and model number for a laptop in the search box. The search engine figures out that you are looking for a laptop and identifies some sites that sell laptop computers.
It takes the information that you provided, and uses the site search on those sites to find pages on the ecommerce sites that match your query. It then presents links to those pages as search results.
That process is detailed in a patent application issued last week which is assigned to ask.com.
Remote execution of actions transparent to a user at registered remote entities in real-time
Invented by Tomasz Imielinski
US Patent Application 20070124289
Published May 31, 2007
Filed: November 30, 2005
Abstract
An apparatus and a method of remotely executing actions in real-time transparent to a user are described. The method includes an execution engine that receives a natural language executable string from a user, identifies a remote entity capable of executing the natural language executable string, sending the executable string to the remote entity to remotely execute an action using information from the executable string, and receiving a result for the user from the remote entity based on the action.
Not only will this process provide informaton from a site search, but it also envisions collecting profile and credit card information from searchers, and filling out information on ecommerce site forms for them. And this could happen without the people searching even seeing the sites that the purchases are being made upon.
A detailed example, from the patent application:
FIG. 5B illustrates a specific example of the method illustrated in FIG. 5A from the perspective of the user interface presented to the user. The user enters a natural language request into the request box 504 of the user interface 502. The request is to “Buy two ticket to the rock concert in New York on October 12.”
The user submits this request to the execution engine by clicking on the execute button 506. The request is then received by the execution engine as a natural language executable string. After identifying appropriate remote entities that are capable of executing the request of two tickets on October 12, the execution engine sends a query. to the user to determine whether the U3 concert on October 12 at Giants Stadium is what the user meant by their request.
The user then has the choice of clicking on a button 510 to indicate that it is the correct concert and to continue, or clicking on a button 512 to indicate that it is the wrong concert.
If it is the wrong concert the execution engine can offer the user interface 502 to allow the user to resubmit a request.
If it is the correct concert, which in this example it is, the user clicks on the “yes, continue” button 510 and the execution engine executes the transaction of purchasing the tickets remotely and transparent to the user. The next user interface 514 presented to the user is a confirmation of the transaction.
In this example the transaction was completed and the information presented to the user in user interface 514 is information on the type of tickets, the price, and the delivery. In an alternate example the confirmation presented to the user may require the user to submit a confirmation to the execution engine to charge the cost of the tickets to a credit card and to specify the delivery of the tickets.
There are a number of other examples, and this patent application has the feel of turning the search engine into a concierge service. Perhaps they retired Jeeves the Butler too soon?
One of the examples listed involves a search to “Find all patents with the inventor T. Imielinski that issued in 1999” and it would search a place like the USPTO for me. I could grow to like performing those kinds of searches on ask.com.
Bill,
How do your Ask bot visits compare to the other search engines?
Hi David,
The ask robot has visited 519 times since January 1st, compared to 5,095 for googlebot and 19,196 for the Google FeedFetcher bot, 79,162 for Yahoo! Slurp and 21,113 for the MSNBot.
Hi Steven,
I was having troubles figuring out what benefit I would get in using a system like this.
Would the sites that Ask.com chose be chosen objectively, or would they have a commercial relationship? If so, would that be disclosed? Would it have to be?
I spent a couple of hours yesterday getting to know Ask’s new interface I like it a lot. I think the way they’re presented all the vertical searches works well.
What you describe here is interesting, though I’m not sure how comfortable I’d be letting Ask be a concierge. Why give yet one more site credit card information. I see the benefit for Ask and I can see the benefit for the end user in that they can manage a lot of things all from one source, but I’d still have concerns giving out sensitive information where it isn’t really necessary.
Now Ask is fetching information from Wikipedia, IMdb and other acclaimed leaders of their fields. I think that when they’ll start offering commercial services in searches, it will also come from leaders.