Conversational Computing and search

From Star Trek’s omniscient LARCS computer to 2001: A Space Odyssey’s slightly creepy HAL conversational computing has been something of a sci-fi fan’s dream. That dream may now be getting closer to becoming a reality. Google’s I/O 2018 gave us an exciting preview of just how seamless and human-like conversational computing can be:

Despite some controversy surrounding that impressive demonstration SEOs and marketers paid close attention because this was the moment I predicted when In Google Semantic Search I wrote how search is heading towards a screenless, keyless future where the main interface will be voice and visual.

Like every watershed moment there are several different types of breakthrough represented here any of which would radically change search, marketing and branding but which, collectively, will change everything forever. I am not given to hyperboles so it’s worth thinking through what exactly has transpired to get us here and then we should be able to see where it is leading us.

Artificial Intelligence and Natural Speed Processing

Natural speech, in order to be understood, requires an enormous amount of real-world knowledge. Ambiguity doesn’t just enter the conversation each time we use synonyms, homonyms or slang. It also enters the conversation each time we interject real-world experience that forms a baseline of understanding of how the world works to quickly convey meaning.

Despite Google’s massive collection of real-world facts and its entity building efforts the real world is in a constant state of flux. The relational values between entities change depending on context and culture which means that the context collapse that’s caused by our technological advances (increased use of socio-technic platforms and constantly-on lines of communication) presents unique challenges when it comes to identifying and processing what’s real and what isn’t within a particular context.

Google is addressing that by increasing the computational capabilities of its platforms, using machine learning algorithms to identify (and sometimes predict) contextual meaning and by adding artificial intelligence (AI) to everything via its Tensorflow cloud AI layer to virtually everything.

There is a real advantage to using a services-wide AI layer to structure and share information. Because an AI has processed it, it is eminently machine-readable and transportable across services. That in itself is an amazing capability to have and it is something few companies outside Google’s unified universe of services can even get near, but the benefits don’t end there. All this processing cross-referencing power would have achieved little if it did not also have a massive amount of raw data to train itself on.

In May 2017 Google announced that there were over 2.3 billion monthly active users on Android devices across the globe. This represents a massive amount of data capturing that, depending upon individual users’ settings, provides Google’s systems with the data necessary to help train their AI. Data is the fuel that runs every AI development and creates personalized, contextually useful touchpoints between users and technology. Google has more of it than anyone else on the planet and that includes Facebook.

Data is the Determining Factor

To illustrate the difference having all this data makes consider that Microsoft whose focus on conversational computing and personal assistants that was kicked off with the Cortana initiative on April 2 2014, has had to partner with Amazon’s Alexa to keep up with Google’s massive leaps and bounds.

And its current strategy of forking out money in an effort to stay relevant in conversational computing reeks of its efforts to match Google on search by partnering with Yahoo and purchasing Powerset to help with semantic search.

Google Holds all The Aces

Personal Assistants, in order to work require data and because they are there to well, as the name suggests, assist us personally, withholding that data from them hobbles the service we want to enjoy without necessarily protecting our privacy.

Because of its dominance of search and its many, now fragmented, iterations Google is in a position to deliver more useful information to more search queries than any other personal assistant. Voice search and, even more to the point conversational computing that delivers music, movies and reminders, sends (and receives) emails and controls IoT devices changes the nature of computing from one which we have to learn to one that has to learn us.

This brings us back to culture, context and context collapse and how it all affects us as individual technology users. Google’s announcement of additional many more languages being added to Google Home by year’s end is the clearest indication we have that the race for indexing, processing and understanding of information is being won by the Mountain View giant.

Here’s an itemized summary of all the breakthroughs the Google I/O demonstration of voice computing represents:

  • Better Artificial Intelligence (AI). Convincing mimicking of human speech patterns in conversation and a pretty good understanding of conversational English.
  • A presence on more home and handheld devices than any other company on the planet.
  • More data than any other company on the planet.
  • The ability to process and analyze all the data they capture (Google has built its own computer chip to do this better).
  • Better voices leading to better human/machine interaction at the conversational computing input point.
  • More services across its universe than just about any other tech company.
  • Better cross-referencing of data from different platforms thanks to its ability to identify entities and enrich its entity graph.
  • Faster expansion of its semantic capabilities as evidenced by the ability to open up the services in different languages; a thing that requires clear contextual understanding of cultural references unique to each language.

So, Is SEO dead?

It’s ridiculous that I have to address this question at every iteration of current information retrieval technology. It shows that we constantly hope that at least one part of our online existence will somehow become instantly easier because our technology has become more complex. It doesn’t make a lot of sense. Nevertheless here’s the answer: No. In order for something that's as necessary to visibility in search as SEO, to disappear either we have to stop using search or search itself needs to change so radically and drastically that it works in an entirely new way. None of this is happening. 

However, SEO is morphing because search is morphing. When conversational computing becomes the primary input format in Google Home and handheld digital devices the nature and context of the search queries changes. But the threshold barrier at the Human-Computer interaction point is also significantly lowered so usage patterns increase. The affective component that ties us to our tech goes up which means loyalty becomes relevant.

These are considerations that show that SEO is expanding to include elements that until recently would have been the exclusive province of marketing and branding. This convergence makes SEO even more complicated. It makes it bigger, with a more complex skillset that now includes everything from technical SEO and on-page optimization to social engineering, human psychology and cultural knowledge.

The insertion of AI and machine learning in information retrieval makes details even more important so SEO activities that are not granular by nature and don't include everything from perfect pictures, great text, some degree of originality, multimedia, relevant cross-linking, link descriptions and everything you can imagine that would make content relevant to an end-user’s search query are missing opportunities to increase visibility on search. Every SEO effort in the past was intended to help project our real-world into machine code. We worked with limitations imposed by the imperfect translation of search queries and the possibilities of their intent. Now machines are entering our world, their limitations dictated by their capability to understand our language and intent. It is a significant transition that in itself will change everything. 

In SEO Help I wrote about activities that create "data density". Now we must pay attention in how we create semantic density. Cross-linking and validating the data trail of our passage through the digital domain in ways that allow machines to fully understand what we do and who we are so that they can provide a business as an answer to a relevant query.