POV: There’s A New Assistant in Town: IBM’s Watson Assistant
Mindshare Point of View
At the THINK 2018 Conference in Las Vegas this week, IBM introduced Watson Assistant, the first non-consumer facing smart assistant. Watson Assistant will enable the first foray into the realm of virtual assistants for enterprises and brands and their products, enabling companies to create branded Assistants for their own products and services without having to give away data to a third-party platform.
Details and Implications:
With no branded device or branded wake-words, IBM has white-labelled the technology behind Watson to create the new offering, allowing companies to completely customise their Assistant to match their own unique tone of voice and brand.
Instead of selling branded devices like Amazon Alexa or Google Home, IBM will be licensing its technology to companies, empowering them to create unique Assistants by inputting their own datasets. Though Watson Assistants are shaped with a company’s private data, the data is localised, meaning brands can’t piggy back off each other’s data but this also mean that brands aren’t giving away their data to Google and Amazon. Privacy is at the forefront of Watson Assistant.
The technology surrounding IBM’s assistant isn’t new. Developers within the IBM Cloud will be familiar with components like Watson Conversation and Watson Virtual Agent, both of which were combined, upgraded and updated to create Watson Assistant. However, when used on devices, the Assistant is completely unrecognizable to the point that users may not even realise an Assistant is there. That’s because IBM has designed its new assistant to be invisible, giving brands full autonomy over their unique Assistants so that they can use Watson to service their products.
One example shown by IBM is a Watson-powered hotel that automatically checks guests in while their room is being set up with their favourite music and temperature preferences and lets guests unlock their room with an electronic key via their mobile devices. Having an open API like Watson Assistant is both good and bad – it is completely customizable and you can build your own integrations but you have to build those integrations yourself.
IBM is not the first company to white label a conversation-based UX platform. Also on the market is Microsoft’s LUIS, a learning-based language service that builds natural language into apps, bots and IoT devices. So the big question is will Watson take the battle away from the branded assistants and bring it to the white-labelled and customisable API world?
Watson is already being used by several brands. IBM and Harman have joined forces to create an interactive and smart dashboard for a Maserati concept car and Watson Assistant has also partnered with Munich Airport, creating a version of the Softbank ‘Pepper’ robot that helps travelers with directions and general inquiries. Other brands using Watson Assistant include: Motel One, KAON and Royal Bank of Scotland.
Watson is probably the world’s best-known AI platform, behind Siri, Alexa and Google Assistant. The fact that IBM has decided to open up Watson to individual brand development may provide a catalyst for a whole new world of brand specific AI usage – which could have implications for all aspects of marketing. The ability to personalize the technology may prove very appealing to larger brands and brand owners who are keen to both maintain their own direct relationship with their consumers in a world of voice and also retain all the data associated with these services. This area will only become more interesting.