Will ‘Personal Assistants’ Eventually Kill ‘Search’?

Siri The short answer is “probably not.” A longer answer is “maybe.”

Here’s a scenario: a natural-language-understanding mobile app that allows users to tap directly into content and conduct transactions (book hotels, make restaurant reservations, buy movie tickets) without the need or intervention of the infamous 10 blue links. Wouldn’t that be generally preferable to the current method of searching and choosing a link and then searching and filtering on the publisher site?

In fact this was the pre-Apple vision for Siri: a tool that could take you all the way from query to transaction, without the need for Google or its search competitors. Apple has veered somewhat from that vision. And though it’s still possible for Siri to relegate mobile search to the margins, Apple isn’t making the kinds of aggressive moves necessary to turn Siri into a true web search alternative.

In fact, Apple has continued to rely heavily on mobile search. As you know, the company uses Google as the default search engine when Siri doesn’t have an answer. In addition, Apple inked a deal with Microsoft to offer Bing results as a kind of “everything else backfill” where the iOS7 version of Siri doesn’t have structured-data — before it resorts to Google.

Google itself is working on “conversational search” and making its mobile and Android search experience much more like a personal assistant. Google Now is a key part of that.

This morning the New York Times published a piece on “predictive search” citing Google Now and listing a number of companies broadly in the segment: Cue, reQall, Donna, Tempo AI, MindMeld (Expect Labs). Some of these companies are striving to be thought of as “personal assistants” and some avoid the label. All of them are trying to use data, “artificial intelligence” and complex algorithms to create a better user experience in their respective domains.

Predictive search and “personal assistant” are not entirely synonymous. A virtual or personal assistant app may use some of the techniques of predictive search to deliver content. However all virtual assistants don’t have “search” as a core piece of functionality.

Tomorrow my colleague Dan Miller has organized a webinar on the state of personal assistants (free, 1:30 Eastern). I’ll be on it in addition to Dan. It also features Norm Winarsky, one of the founders of Siri and a current VP at SRI Ventures, which has produced several personal assistant apps/companies post-Siri. It also features Phil Gray of Interactions Corporation. We’ll talk about what defines a personal assistant, current applications of the technology and future opportunities — for consumers and enterprises.

A number of companies in the market have invoked the “personal assistant” metaphor as a way to separate from search (and Google) or to take advantage of the hype around Siri, which has now died down substantially. In some cases “personal assistant” is simply a marketing term but in other cases it refers to a set of core technical capabilities and functions that truly advance the user experience.

For example, we spoke today with Ask Ziggy, another very interesting company offering natural language understanding and personal assistant capabilities to third parties and enterprises. Nuance has a similar enterprise offering in Nina.

Eventually, at least in mobile, “personal assistants” or “virtual assistants” (PVA/VPA) will likely replace many of the functions and use cases currently involving search on mobile devices. Search will never be obsolete but over time it will be a lot less frequently used than it is today.

My view is that there’s also an opportunity to migrate a solid mobile “assistant” experience onto the PC — or at least the tablet, which is supplanting the PC in many instances.

If you’re at all interested in the topic of personal assistants and where they might be heading, tune in tomorrow to the webinar (pre registration required). The webcast will be slightly enterprise-centric.

But I’m also interested in your thoughts below about the longer-term mobile and tablet user experience. Do you think that search will be partly replaced by these “assistant” apps or functions (including from Google)? Or do you believe that search and mobile search will continue to live on much in its current form — undisrupted?

You can follow any responses to this entry through the comments feed.

4 Responses to “Will ‘Personal Assistants’ Eventually Kill ‘Search’?”

  1. Yishai Knobel | HelpAround.co says at

    No, Personal Assistants will not replace traditional Search.  But something else will: Social Graph Search.  Millions of consumers already turn to their social networks to get help and advice;  it’s just not easy enough. yet.

    Our bet at HelpAround.co is that, if made easy enough, consumers will prefer to get help from friends instead of from Google.  It’s less work for them, and MUCH more fun.

  2. Greg Sterling says at

    The social graph is being incorporated into traditional search as a ranking factor or variable. In terms of Facebook, it remains to be seen how successful the company’s search efforts become.

  3. Raj Singh says at

    Hi Greg,

    I have a couple comments:

    1. I don’t think of assistant as a category. The technical definition of an assistant:
    a. Something that learns
    b. Something that completes tasks for you

    Just like how LBS used to be a category but now is a layer, all applications can benefit from being more assistant-like or anticipatory and we are seeing this across applications. For example, FourSquare upon opening will suggest lunch places if it’s lunch time. It’s using context to be more anticipatory. 

    I like to think of the present time as the anticipatory web (vs Web 2.0, the participatory web). It’s the use of data and context to deliver a more personalized and/or anticipatory experience…

    2. I don’t think of assistants as horizontally as you. Google Now is an example of a very horizontal assistant (and they can achieve that b/c they have access to your search history). But I suspect most assistant functionalities will actually be quite a bit more vertical and so they may affect some search marketshare in a particular vertical (eg the Amazon app gets very smart about suggesting products you need in mobile context) but I don’t see one company horizontally achieving this. It’s too hard and there is much too much data causing noise (signal to noise ratio). Most horizontal assistants, IMO, will remain broad and shallow (like Google Now) and these assistants may take some search marketshare but arguably the search results that the users would have made would have yielded OneSearch type of results anyways (meaning, they are Siri-like searches where there is a defined answer).


  4. Greg Sterling says at

    Good points Raj. I agree with almost everything you say and agree that verticals will probably be the place where this functionality takes hold first. I do think that there will be horizontal assistants. In particular the carriers have an opportunity but I’m not sure that they’re capable of getting the user experience right. 

Leave a Reply