“Reference Resolution” was cited at Google I/O and it’s directly connected to owning your brand in voice & customer experience
Scott Huffman (VP Engineering, Google Assistant) walked the Google I/O community through some new advancements with Google Assistant this week.
There were exciting moments in the demo where his colleague, Meggy, executed not only cross application (dropping a photo into a text thread, using your voice, toggling appropriately from app to app) but also an impressive display of command speed (back to back to back to back commands executed with near zero latency, definitely less latency than click and swipe controls).
There was a moment when Scott walked us through an individual use case that named a function that is at the core to a voice and virtual assistant.
The use case was his desire to ask his assistant how long it would take to get to his mom’s house. The nuance being, there are lots of places in the world that are called mom’s house, not just HIS mom’s house.
Because his contacts were accessible to Google Assistant (and this is concept that is key) Google Assistant is able to take his personal reference and prioritize it above the restaurant called “Mom’s House” in Korea Town, Los Angeles, for instance.
https://www.youtube.com/watch?v=TQSaPsKHPqs
Minute 29:00 – 30:46
The key element in this use case is access to data. The user (Scott) allowed Google Assistant to leverage his contacts to make informed conclusions. So too are there options to allow Google Assistant access to email, calendar and other services.
Access to these data points gives way to better ‘Reference Resolution,’ a linguistics term that informs language interpretation.
‘Reference resolution’ is what it sounds like: the ability to understand what is referenced in language, accurately. There is no shortage of elements that influence language: spatial context, location, recent conversation, people around us ,and the list goes on.
These elements are data. As humans we process this data with all of our senses, in real time, using intelligence.
Voice Assistants too have intelligence. This is where the idea of voice assistants gets confused for many professionals.
Voice Assistants don’t promise to automate the existing customer experiences (web, mobile, native, call center, product support, employment applications, appointment making) you have created into a conversational experience.
Voice Assistants promise to leverage the data you offer them in a meaningful conversational structure, to deliver content, features, functions and experience to those people who want to engage with you.
As an individual user I can’t expect my assistant to tell me how long it will take to get to my mother’s house unless I have told it (or authorized it) to know where my mother lives. This should seem logical. Albeit in the ‘vision of the future’ we are given I can see people assuming that Google should ‘just know’ where my mother lives; it isn’t actually the case and we should all hope it never is.
Brands Own the Data Leading to Reference Resolution
In that same vein, brands can’t reasonably expect that voice assistants should ‘just know’ how to communicate about something that is brand related. And brands should be rejoicing and embracing that fact.
Just as end users we decide if we want our assistants accessing our contacts and email, so too brands are deciding (often by absence) whether assistants can access the conversation about their brand.
You don’t want Google and Amazon choosing what to say about your brand when questions arise, and interest is expressed. You want them connecting the dots from other pieces of context they have access to, in order to leverage your voice experience correctly.
But letting the assistants determine what they say about you is exactly what is happening for many brands in today’s marketplace. Amazon has a relationship with Bing and widely references Wikipedia and Wikidata for its responses. Google uses its own portfolio of web indexing on its platform.
This content was built for web browsing, clicking, swiping and pinching – depending on the device it was designed for – but not voice.
This is resulting in awkward and irrelevant responses by the assistants when brand or subject matter specific inquiries, transactions and tasks are given by users.
Just as companies, governments and individuals began leveraging websites in the early days of the internet in simple yet fundamental ways, so too must they in voice. First iterations of voice experience will be training the business as it meets basic expectations of the user. But soon more complex models will evolve, and given the pace of adoption the voice channel is set to overtake click and swipe, in a meaningful way.
Laying Groundwork for Tomorrows Promise
The creation and management of your own voice app means that when I decide to give my wife a pair of shoes as a gift and she doesn’t like them (this would never happen mind you), I can ask something as general as ‘How do I return these shoes?’
Google Assistant may reference my email, find the receipt for shoes I bought last week and drop me into the voice app for “Shoe Brand” where I begin the process of requesting a return slip and have it texted or emailed to me so I can print it out.
The assistant is offering value and making things easier by connecting the data and capabilities it has access to, including the content and function of the brands voice app.
Having conversational content of meaning for the assistant to access is a critical point that brands need to recognize and adopt, quickly. Not only to meet the expectations of customers but to create a competitive moat for themselves.
This use case isn’t far off, albeit more sophisticated than today’s capabilities. What is possible today is entering the channel and preparing for the already prolific use of voice by human beings globally.
Post Script
It’s for reasons like speed to market, brand ownership, customer experience and extensibility that Voicify was created. We believe when companies offer support, problem solving and functionality through voice channels both they and the customer win.
Voicify gets brands into voice in days and weeks, not months. No matter what the motivation, Voicify is the enterprise solution that solves the voice challenge.