With the latest wave of artificial intelligence brings the latest wave of fears about job displacement. Encouragingly, previous technology advancement has shown us that there are always opportunities for people to provide value to the business in other ways once activities become automated.
I, for one, welcome progress in automating tasks. In our culture, it’s common to pride ourselves on how busy we are. “How’s work?”, “Oh, it’s good thanks, I’m really busy”. Being overwhelmed with work, squeezing as many tasks from the to-do list into the working day, makes us feel productive. It’s an image that is re-affirmed socially as well.
“The latest developments in machine learning mean that we have the opportunity to automate more tasks than ever before”
At the same time, if I was to ask you what was needed in order to make improvements to your business or role, many of us would be able to reel off a list of answers or would know how to find out. The largest barrier to progress is time, not knowledge. There are only two ways to solve this; work more hours or do less work. If we assume the second is the only viable option, this will either require de-prioritisation of tasks that do not provide sufficient value or, offloading the task to a computer.
Computers have been proliferating the workplace for years. The latest developments in machine learning mean that we have the opportunity to automate more tasks than ever before.
Indubitably, Watson
Over the last few years, IBM has been investing heavily in this space and this has culminated in IBM Watson. Watson is a technology platform that uses natural language processing and machine learning to make sense of unstructured data. Something that traditional only people could do. You can see a bit more about Watson in the video below:
Power your business with Watson
So what opportunities are out there? Here are some example use cases for Watson and machine learning.
Chat support agent
You can train an automated support agent on your frequently asked questions and it can then handle first line support for your companies products and services. This can be achieved with the Conversation and/or the Dialog APIs. If the agent isn’t confident in answering then the request can be passed on to a person.
Smarter search service
With Retrieve and Rank, you can create a more intelligent search that re-orders results based on “signals in the data”. This could be trained by a domain expert choosing the best answers for a set of questions. The system can then find trends in the data to handle new questions that it hasn’t seen before.
Better customer segmentation
If you want to achieve more targetted segmentation of customers or match people with other people or companies, then the personality insights API is for you.
When the service is provided with content written by an individual it can paint a picture of the personality traits it believes they possess. Naturally, the more words, the better the accuracy, but it recommends a minimum of 3500 words.
Use cases for this range from customer acquisition to customer marketing and customer care. At each step, providing a more personalised service without it needing to be done by an individual.
“[Automatic] image recognition accuracy has surpassed that of people so this has become a real opportunity”
Image recognition
Many tasks require manual visual review. This is a prime opportunity for more automation if a machine can be trained to process these tasks. In recent years image recognition accuracy has surpassed that of people so this has become a real opportunity. The Visual Recognition service provides a mechanism to create your own classifiers that you can train to recognise objects.
This could lead to faster insurance claim processing through being able to automatically review accident photos. Or ecommerce product discovery tools that take a photo of an item and then search to provide results of where it can be purchased from.
Resolve conflicting goals
If your business regularly requires the balancing of conflicting goals, decisions can often only be made by domain experts. With trade-off analytics, objectives can be provided to the system and it will then will then optimise to balance between them to get the best results possible. No more hunches required.
Other utilities
Bulk document conversion can also be achieved with the Watson APIs, which is useful if you have documents in PDF format that are needed in plain text or HTML.
Watson has a language translation service that takes into consideration the concepts discussion in the document in order to improve the accuracy.
Take a browse of the service catalogue and review your business processes to identify if there are any opportunities for automation. Replace those and you can prioritise your time on your business development goals rather than spending your time running the day to day.
Success isn’t guaranteed but perhaps the Research and Development tax relief can be used to make it a more attractive business investment.
