3 Key Ways AI Is Changing Data Analysis in 2021
Artificial Intelligence has been predicted to be the big change to our lives in the 2020s and without a doubt, the changes will be widespread and rapid.
In fact, in a recent survey, PWC found that the vast majority of executives have either looked at or started implementing some form of ML or AI in their businesses.
This means that the amount and depth of data will be increasing rapidly over the next few years as will the requirement for fast and accurate data analysis. So, what changes will we see in data analysis in the next 12 months and what opportunities will they present for connected companies?
What are AI and ML?
Often people will use Artificial Intelligence (AI) and Machine Learning (ML) interchangeably and this is understandable as there is a fair amount of overlap between the two.
Essentially the difference between the two is the way that they cope with unexpected events.
For example, if you tell a machine learning program to search for pictures of cats then it will do this quicker and more efficiently than a human. But when it comes up with a picture of a dog, it won’t know what to do with it unless specifically instructed.
With artificial intelligence, the program will search for pictures of cats and then when it finds a dog, will search any connected information sources to find out what it is. Only when it finds an insoluble problem will it need human intervention.
But the potential for confusion is clear and it’s certainly not helped by unscrupulous vendors muddying the waters even more!
Integration of IoT and analytics
The development of advanced internet connectivity with improved WiFi, 4G and 5G networks, and extended coverage of space technology such as Elon Musk’s Starlink mean that Internet of Things (IoT) connected devices are now more viable than ever.
Being able to connect sensors embedded in devices as diverse as cars, aeroplanes and washing machines means that manufacturers and service organisations have access to usage and fault data at a rate never before seen.
The sheer volume of data produced has even resulted in the term ‘Big Data’ being coined to describe a collection of information with both depth and breadth and more importantly, real-time ability.
But herein lies a problem; this massive volume of data means that for most companies, their IoT-enabled machines are producing information that is so extensive that traditional analytics techniques simply won’t work.
Often, the skills required to carry out this work are in short supply and the time taken to produce innovative and insightful analysis means that the real-time advantages are lost.
This is where AI and ML come in.
Using well-developed AI systems means that huge volumes of information can be processed very rapidly indeed leading to quicker analysis and insight.
What is interesting is that this has now morphed from a purely internal analytics capability to an external customer-facing marketing focus.
For example, several retailers are now utilising IoT-based technology to serve advertisements that are specifically targeted at individual customers. When the customer visits a store they are tracked using IoT-enabled cameras and as they approach specific displays, advertisements are shown that are tailored to their specific interests.
When the customer buys, the information gained from POS systems is then fed back into the data pool to enhance the picture of their preferences and in turn, further refine future marketing.
The data processing speed required to achieve this can only be achieved using forms of ML and AI.
Increased conversational analytics
There can’t be many people who haven’t visited a website with chatbot functionality.
A well-programmed chatbot can increase customer satisfaction, speed up service response times and reduce costs for businesses that choose to adopt them, so they make complete sense.
In the past though, these have been fairly crude devices with a small number of pre-programmed responses to pre-predicted questions.
The problem is of course that people aren’t pre-programmed and so the questions they ask are highly variable.
It is also fair to say that keeping a chatbot up to date with potential questions is an onerous task and when the business is vibrant and growing, this could end up being an impossible job.
There is also the issue of mechanical responses being given that make it clear that they aren’t speaking with a real human.
However, the advances in AI with respect to chatbots means that they are now much more capable tools with up to 80% of questions now being able to be answered correctly.
That’s not to say that there haven’t been problems along the way, but in general terms, the advances in AI-enabled chatbots have been remarkable.
What is more interesting is the human-like interactivity of these devices with naturalistic language and the adoption of different tones of speech that mimic the person with whom they are conversing.
From a data analysis point of view, being able to not only respond to queries but also ask questions of the customer, means that the AI system can be used to prompt information gathering.
The types of questions people ask could be used to highlight specific failures with a product, predict potential marketing opportunities or develop new support features but this is only possible if the information gathered is correctly analysed.
Greater commercialisation of AI and ML
As with many technological breakthroughs, the initial days of AI were dominated by academic research at universities and technological institutes.
However, in the last few years, the technology has developed to such a degree that the vast majority of the activity in the ML and AI sphere is by commercial businesses.
In some cases, these are for purely internal usage, such as businesses that seek to collect and interpret data from IoT-connected devices that they have supplied or serviced.
But in many others, AI is being supplied as a service to provide commercial functionality.
For example, systems giant SAP developed HANA as a data analysis tool which allowed Walmart to take information from 11,000 stores across the US and interpret it within seconds.
But the commercialisation of AI isn’t confined to global companies with deep pockets.
In actual fact, the internet is awash with businesses that are developing AI and ML systems to help with any number of issues that both business and consumers face.
The type of application ranges from data analysis dashboards to allow better decision making, to sales channel optimisation services or even failure prediction applications for heavy industry.
This produces something of a ‘virtuous circle’ with more companies adopting AI for data analysis simply because there are more useful commercial applications available. The more companies that invest in advanced analysis systems, the more it becomes acceptable in their particular sector.
AI & ML are coming to your company
The plain fact is that the development of AI and ML means that there won’t be any business that is unaffected.
This means that in many cases businesses will be forced to adopt AI systems simply to keep up with the competition, but it also means that companies that adopt early can steal a march on their rivals.
AI can help with operational tasks, but the real value is in analysing the information that is produced by all of these interconnected systems and coming up with insight that will make a real difference to the outlook for the business.