Bhumika Shrestham answers questions through an AI interface

Janina Gavankar is not a name that readily brings up a famous face but mention Ms Dewey, Microsoft’s gimmicky and now retired search engine project and a great many people know who we’re talking about.

As the embodiment of Microsoft’s search at the time, Janina, used over 600 videotaped snippets to comment on search queries, searcher’s supposed habits and respond to some of their questions. She was sassy, bold, playful and occasionally risqué. What made many love her more than the plain vanilla Microsoft search was the fact that in the interaction she occasionally managed to get her pre-recorded snippets just right in relation to a search query and it felt like magic. Search, the usually invisible, mysterious interface that helped us navigate the world better, suddenly acquired a human face. 

Ms Dewey was a Microsoft Marketing Project for search
Ms Dewey's sassy attitude was a big attraction for Microsoft's search engine project

Now meet Bhumika Shrestham the first Nepali to have “other” gender passport but also the face of a sophisticated AI that has a background personality that allows it to sample the world of information through Shrestham’s personality filter and return answers to search queries related to her experiences and transgender issues in Nepal. 

Bhumika Shrestham AI search engine
Bhumika Shrestham combines voice and rich information snippets for as thorough a search engine experience as possible

This is one more of Asia Week's Red Letter Projects and when compared to Ms Dewey this is an upgrade of several orders of magnitude and it also points to the future of search. The web may be global but its impact is local, though the concept of localization is now applied to informational as well as landscape topology. Information becomes valuable only when it becomes personal or at least personally applicable. To achieve this in a world where data increases exponentially by the day, search has to become not broad but deep and it has to become personable as well as personal. 

Google understood this very early in the game by putting in place filters that aimed to personalize search through context and location. The breakthroughs that have been achieved of late in machine learning and artificial intelligence are now making it possible to create a multi-faceted search experience that transforms data from informational to transactional. In the process it also personalizes the contact between the searcher and the search engine interface creating a relationship that is truly symbiotic: Without a personalized, deep search engine experience we are each downgraded to the biologic limitations of the processing power and speed and information storage of what’s inside our skulls. Without deep, nuanced and detailed access to our personal data and usage patterns search, no matter how intelligent it might be, is essentially lobotomized. We’re back to the hit-and-miss Ms Dewey world.

The Asian Week Shrestham project is powered by IBM’s Watson and a few more machine learning driven tools. Google, IBM, Amazon, and to a lesser extent Apple and Facebook are now engaged in a battle to commoditize machine learning and usher an age of AI-driven apps and tools that reside in small-form devices (phones and tablets). This is happening now. 

Marketers, branding specialists and search engine optimizers are challenged by artificial intelligence. The questions that they ask are usually along the lines of “what are the shortcuts available to us now?”, “how do we scale our efforts successfully?” To truly understand what needs to be done in this smarter digital world we need to see what the effects are at the point where the technology meets those who use it. 

AI interfaces invariably provide: 

  • Personalized responses
  • Deeply relevant, localized responses
  • Results and information that are richly detailed
  • Results and information that can be quickly veracity-checked
  • Information that is filtered through the end-user’s intent and purpose
  • Results and information that are thought to be of high quality and accuracy

Reverse engineer this approach and you end up with a list of tasks that feed into a granular breadcrumb trail which leads to a company’s, person’s or brand’s digital identity. In the simplest word possible, entities.  

You also need to, however, have a social media impact. Pretty much the same way you impact upon people in your immediate vicinity in the real world whose perceptions of you and your importance (and skills) then go into the building of your reputation. Reputation feeds into the calculations that determine how much trust you have earned and how much trust those who don’t know you are willing to give you. Without trust you won’t get anywhere.

And just like that we’ve gone from looking at hard requirements to soft ones. Our technology, these days, is particularly adept at stripping back the layers of anonymity and facelessness that we’d become accustomed to as the norm over the last century and surfacing the things we need in order to humanize our interactions at scale. 

For a detailed, step-by-step guide to the practical aspects of semantic search, marketing and branding you should check out SEO Help