The Voice Assistant Breakthrough
With over 22 million combined units sold for Amazon’s Echo and Google Home, ‘smart speakers’ have certainly been a breakthrough in AI for the mainstream. These voice speakers are now quite adept at understanding human speech and executing commands to activate the arsenal of apps on our devices to do anything from looking up weather to playing song
Computers with Human Intelligence – Not Yet
However, if you ask the world’s AI experts like Geoffrey Hinton from Google, Yann Lecun from Facebook, or Demis Hassabis from DeepMind, there is a consensus that we have a long way to go to create what’s called Artificial General Intelligence which replicates human cognitive abilities.
The closest example that we have to-date is the breakthrough that occurred with Deepmind’s AlphaGo computer system being able to soundly defeat several of the world Go champions in an extremely complex and strategic game. The good news is we still have some upside runway for what today’s AI can do for us.
Beyond Smart Speakers to Intelligent Digital Assistants
For businesses there is an opportunity to build more intelligent digital assistants to interact with their customers in completely new ways. However, for this to happen, there are a few critical facets that will truly make their digital assistants ‘intelligent’ from the eyes and ears of their customers.
- Access to Customer Interaction Data for Personalization – most companies are maintaining large volumes of transactional data. With AI’s new natural language abilities, untapped unstructured data like call recordings, chat and emails can also be deciphered to tailor customer conversations based on what they’ve told you they’re interested in.
- Connection to the Knowledge Graph to Make Proactive Suggestions – advanced analytics and machine learning can be used to create forward-looking insights about customers such as their propensity to buy certain products, their likelihood to attrite and even how each customer prefers to transact (i.e. what channel and what device). These insights can be used by the digital assistant to use to make proactive suggestions for each customer, much like a human personal shopper might make.
- Learn and Remember Each Customer Uniquely to become a True ‘Personal’ Assistant – this might be the most complex component of intelligence that a digital assistant can be built to perform. It takes a completely different back-end configuration of the AI assistant, but it will differentiate each company that make this kind of investment in their customers.
- Ability to Seamlessly Collect Elusive Customer Preferences and Styles –an intelligent digital assistant can be trained to seamlessly ask questions such as style preferences, current needs, sizes and even who they buy gifts for. In return, they can provide back to the customer a suggestion that helps with whatever they need at that time. An example of that is luxury brand 24serves whose trained virtual fashion stylist gathers your fashion profile and provides personalized outfit recommendations.
- All the Best Execution Features of Alexa, Assist and Siri – at a bare minimum, the speech recognition and online information retrieval will be expected. In the case where a company chooses to have their digital assistant work through a smart speaker or messenger, this capability will be taken care of.
- The Right Persona for the Right Customer – a very underrated component of a digital assistant is the level of assurance, and conversational tone conveyed during the interactions. Defining the right Persona should first consider the specific circumstance in which a customer would be conversing with it. Companies can create multiple personas that vary based on the profile of the customer so that ultimately customers will build a trustful connection with the assistants.
With the press of an icon, a Digital Assistant can be instantly answering a customer’s questions, seamlessly learning more about them and finding ways to proactively delight them with personalized suggestions and offers. Don’t throw your Echo away yet, but this added intelligence is the differentiator between a voice assistant that retrieves information and a personal assistant that is personalized, proactive and learns to get customers what they need when they need it.