One of my favourite poems growing up was “If all the world were paper“, originally penned by John Mennes and James Smiths in 1658, because it made me THINK. It made me question the status quo. What if the world was paper? What if the seas were ink? What if indeed! For a child, the imagination ran wild. My grown up version takes on a slightly different perspective…
If all the world were digital
If all the tech in sync
If all the data
Were sliced and diced
How could we outthink?
In a world where all businesses are digital and make data-driven decisions, how do you create and sustain a competitive advantage? Businesses today need to intrinsically understand, reason and learn – that is the foundation of cognitive.
With all the media and excitement around Watson (including from me!) there has been a fundamental misconception about what Watson really is. Here’s the crux of it: Cognitive is not Watson. But Watson is most definitely the greatest example of Cognitive technology today!
IBM Watson was born out of a challenge to create a computer that could understand natural human language, investigate, reason and learn. Set the enormous task of beating the world’s best Jeopardy champions, and like all transformative innovations, it failed miserably in its first attempts. But as the technology evolved, Watson learned the nuances of the human language – including how to understand questions and structure answers. The phrase “the cat sat on the mat” is not just a collection of words – they have meaning – and Watson understands the inference that the cat is on top of the mat. With the ability to read and understand, Watson then learned to investigate and to reason. With the world’s information at its fingertips, Watson can trawl through unfathomable amounts of unstructured data – wiki pages, medical journals, clinical notes, maintenance records, legal documents – and make inferences about which information is most relevant to the question at hand, draw conclusions about what the answer is most likely to be, and recommend an answer based on confidence levels.
Most importantly, Watson can learn. When a question is answered incorrectly, Watson incorporates that knowledge into its corpus to improve the ability to answer similar questions next time.
Watson has since been productionised and put to real work – helping clinicians diagnose and treat cancer patients, helping war veterans deal with life after the army, helping employees make critical decisions based on lifetimes of expertise, just to name a few. Make no mistake – Watson will change the world. And yet I still get asked the question, what is the different between Watson and a search engine like Google?
Here’s the simplest way to describe Watson:
Imagine you wanted to buy a car. Most likely you would first get on the Internet and search for cars. You’d collate information about dealers, models, styles, features, pricing, financing, formal reviews, social comments, and you’d work to read and understand in order to be able to make an informed decision about what is best for you. In the scenario of buying a car, the amount of information is consumable – but keep in mind, that is not the case in the world of healthcare or law!
Google does step one. I’m looking for a “awesome raspberry red seven seater SUV” – a Web search would return web pages that contained a high number of “awesome“, “raspberry“, “red“, “seven“, “seater” and “SUV” with no real understanding of what the words mean in context. Articles about an “awesome blue five seater SUV” would rank highly, even though it’s not what I want.
Watson, on the other hand, mimics the entire human and information interaction, at epic scale. Watson can consume incredible volumes of information from web sites, manufacturer documents, maintenance records, and social media. It can understand that the colour of the car I want is not just red, but raspberry red. It understands the SUV needs to have seven seats. And it can return options to me based on whether people like me thought it was something like awesome. It doesn’t just give me a recommendation, but also the evidence to support the decision so I can make an informed decision myself.
And when it gets it wrong, the next time someone asks for an awesome raspberry red seven seater SUV, it will provide a different recommendation based on what it learned from me.
That is the power of Cognitive.
Of course, asking Watson to help me buy a car is kind of like asking a doctor to put on a bandaid! But you get my point. 🙂
What many people don’t know, is that cognitive capabilities can be found across a range of technologies that have been given the ability to reason, act and learn.
Take for example the case for Next Best Action, which can be used to replicate and scale the human thought process of a highly experienced sales person in a cross-sell/up-sell situation. The solution will reason based on a set of business rules and patterns/trends that had been uncovered using predictive and advanced analytics, it will make a recommendation based on the optimal outcome of a cross-sell/up-sell, and if configured to, it will act – making an offer to a customer or recommendation to a call centre agent. Most importantly, it will make note of the outcome and learn from the result. If the customer didn’t take up the offer – why? The outcome changes the prediction and action next time based on the ability to learn over time – just like an experienced sales person would.
Another example is the cognitive capabilities that have been built into Watson Analytics, designed to replicate and scale the ability of a human to analyse data. Imagine you had your own personal analyst – you’d give them data, they’d analyse it for you and make recommendations on insights they believe you’d be interested in. You would ask questions of them, they’d give you the answers, and they’d learn what you were interested in so next time they give you what you need in the first instance. Think of Watson Analytics as like your own personal analyst. Just like a personal analyst, it will take your data and uncover insights it believes you are most interested in, presenting them in a way that best communicates that insight. You can ask business questions such as “How did my sales track year on year?” – Watson Analytics understands natural language and translates it into a chart that answers your question. And it learns – as you interact (and don’t interact) with certain insights, it better understands what you’re interested in and uses that to serve you, and your colleagues, better next time.
The industry has many definitions for “cognitive computing“, “artificial intelligence” and “machine learning” – and no doubt many would have differing views to mine. But fundamentally, we all agree we are entering an era where cognitive technologies are going to change the world of business.
A cognitive business is a business that thinks. That reasons and learns. It is an organisation that taps into all data – both structured and unstructured, both internal and external, both at-rest and in-flight. It is a company that uses technology to distil the 2.5 quintillion bytes of data we create every day, down to a recommendation that a customer, an employee, a citizen can use to change their world.
A cognitive business can:
– Understand context and aspects of personality to personalise and deepen engagement with customers.
– Collate the most-advanced knowledge available and bring it to every employee to elevate the collective level of expertise.
– Interact with customers to create products and deliver services that continually learn and improve.
– Spot patterns in both traditional data sets and unstructured data to accelerate high-stakes research and time-to-market.
– Leverage vast quantities of both unstructured and structured data to continually improve its processes and decision-making.
So how do you get started on your journey to cognitive?
1. Design a cognitive strategy: Look at your products, services, processes and operations and determine which would benefit the most from the ability to reason and learn.
2. Extend cognitive with analytics: Understanding data is key. Make sure you can collect and curate the right data – structured and unstructured.
3. Move to a cognitive cloud: Make sure your business can get everything possible out of your cloud services, your data sets and your cognitive services.
4. Build a cognitive infrastructure: Businesses today need an IT infrastructure designed for cognitive workloads. They need to be able to handle the data and analytics required by cognitive services.
5. Adopt security for cognitive business: When everything is connected, everything is vulnerable. Make sure everything you do, every bit of data, every transaction is secure.
And do your research. Anyone can add the word “cognitive” to their product or service, but not every product or service has the ability to reason and learn. Only IBM can give you the power of IBM Watson.
If all your world is digital. If all your tech were in sync. If all your data is sliced and diced. How will you outthink…..your challenges? Your competitors? Your limits?
Welcome to the era of cognitive business.