Peter Haggar is up to speak about Watson [JJ: incidentally my favourite bit of IBM tech].
For those who don’t know, Watson is the kit that competed in the TV program Jeopardy, four and a half years ago.
Haggar says Watson enhances, scales and accelerates connections between humans and data. He describes the movement from the tabulating systems era from 1900, to the programmable systems era from the 1950s to the cognitive systems era from 2011.
Data growth is astronomical, yet 80% of data is unstructured. So making sense of available data is difficult – hence the big data era. But while we can take advantage of data to generate business intelligence, it’s still difficult to process data in such a way as to generate deeper insights. Cognitive systems – true artificial intelligence – specialises in unstructured data and never stops learning from the data it encounters.
This means that true actionable insights can be developed. Right now 350 partners (in 36 countries) are accessing Watson insights through Bluemix, and they are using cognitive APIs and services to make the data they collect useful.
Some of the services inside Watson include:
- Personality Insights – takes unstructured text from a blog, a web page or a social media account, and produces a personality profile on the basis of that text-based data.
- AlchemyVision – analyzes images and videos to understand their context, and produces metadata tags on the basis of the content of the image or video. It uses facial recognition (for celebrities) and it identifies gender, age and environment in which people appear – including identifying places by landmarks, etc.
- AlchemyLanguage – takes unstructured text to find sentiment and concepts in text.
- Tone analyzer – helps determine emotions, values and social propensities in written communications.
- Tradeoff Analytics – which helps make better purchase decisions based on multiple competing goals.
- Relationship Extraction – takes unstructured text and finds the grammar components and objectives of those written communications.
- Text-to-speech and Speech-to-text
- Concept Insights – creates implicit and explicit links between content and other documentation
- Concept Expansion – analyses text and learns similar terms based on context
- Language translation.
- AlchemyData News – uses natural language processing to spot trends and track brand mentions.
- Natural Language Classifier – uses best match classes to work out if text messages are in the context of workplace, personal communications, etc.
- Dialog – conversations with bots
- Q&A – direct querying to deal with specific questions
Working together, these tools can help improve the actionable insights to be derived from social engagement, traditional business communication and in research communication. The opportunity is to simplify the information filtration process so that discovery of relationships, opportunities for process improvements and solutions to problems can be accessed more easily.
Haggar uses the example of a customer support call, which uses speech to text tools to generate customer profiles that analyse tone, personality insights and customer objectives to determine the best product fit for a client. This takes some of the guess work out of call centres as well as streamlining the sales process.
Haggar gives another example of a service support app, where a system learns the typical services most needed by an individual client, so the steps to access information are more contextually presented. He also gives an example of insurance claims based on information provided by the client, and by trends in claims.
Haggar says cognitive computing enables deeper human engagement and better decison making. The era of cognitive computing is here. But if you don’t use this tech, it’s going to waste. What we need to think about now, is how best to disrupt the way data can be capitalised.