If you haven’t tried Siri, S-Voice, Cortana or Nina recently, you’re really missing-out. Speech recognition has reached a level where it’s almost uncanny in its ability to convert speech to text, and natural language understanding is likewise advancing rapidly, so machines can now understand a great deal of what we say and what we mean when we use everyday language.
Along with advancing conversational user interfaces, the flexibility that many organisations now have in creating API’s to their internal systems allows Intelligent Virtual Agents to not only understand you and answer questions, but help you transact and simplify everyday tasks. It was only a few years ago that suggesting ‘we’ll build an API’ was met with ironic laughter in many organisations. This trend towards more flexible API’s and easier data access will only increase over time, further improving that capabilities of Intelligent Virtual Assistants.
There are still some awful examples of chatbots out there, but true Intelligent Virtual Assistant solutions such as Nina are pushing the market forwards with high-quality conversational user interfaces and connected capabilities that deliver a provable, positive ROI to large and medium companies and an improved and simplified user experience for consumers.
At the other end of the spectrum there are numerous ‘science projects’ being undertaken by organisations dazzled by the vision of ‘Artificial Intelligence’, by which they generally mean Unsupervised Machine Learning (UML). These capabilities will almost certainly be commercially deployable in the next 5-10 years, but not as a ‘big bang’ launch, and not as quickly as todays hype might make you believe. The fragmentation of an organisations knowledge, located in FAQ’s, knowledge bases, websites, manuals, intranets, staff hard-drives and agents heads, with often vague or contradictory information, makes adopting UML all the more difficult.
For the foreseeable future Machine Learning will be an enhancement to the Designed Intelligence of today’s Virtual Assistants, supplementing their capabilities and possibly using agents to train the system over months or years. Only then will AI gradually take-over more of the Designed capability.
It’s also important to note that AI is often presented as a single ‘thing’, but the reality is that AI is made-up of dozens of capabilities, including speech recognition, natural language understanding, text-to-speech, machine learning, sentiment analysis, text recognition, image recognition etc. etc. The market is made-up of companies from IBM, Apple, Microsoft and Nuance down to small start-ups that focus on a niche area. We’re already seeing acquisitions to acquire specific capabilities, but no single company will be able to lead in all these areas. There will be a marketplace of AI capabilities that will need to interface to create the capabilities a particular organisation needs.
There are some really exciting launches coming-up over the next few months, and the market for Intelligent Virtual Assistants like Nina is growing rapidly, but we really are just at the beginning of the journey.
Wired has a great article about the state of the art here. Note the graphic showing teenagers voice usage to complete tasks on their phones – we are at the very beginning of a major shift in computing that will change how we use consumer devices and how we interact with organisations.