AI is a ‘zero-sum game’

Technology VS Data 

Not all businesses have started capitalising on the benefits of AI, though. A recent IBM survey of over 4.5 thousand senior decision makers found that 22% of respondents are not currently using or exploring the use of AI. For such companies, things must change rapidly. What many people don’t yet realise about Machine Learning (ML) in business is that it’s a zero-sum game. Falling behind on the AI train could make it impossible to catch up. The reason why is not necessarily the AI technology itself, but rather, the key ingredient that makes AI and ML work in the first place: data.

Until now, when it came to innovation, businesses have focused on having the necessary resources to either buy or develop the best technology available that would place them ahead of their competitors. When robotic process automation (RPA) first came about, for example, American companies started falling behind because the Japanese developed better technology. However, they caught up by investing in better software, to level the playing field. It didn’t matter so much who started using RPA first, but rather, which companies were able to catch up with their competitors over time by employing more powerful technology. With regular technology, businesses have the option of gaining the upper hand over their competition if their resources allow it, by buying or developing similarly. 

AI is a very different story because most ML algorithms are open source technology. By sharing the codes and algorithms they use with the open source community, developers allow others to understand the methods they use to manage databases, detect patterns and promote machine learning, and in return, they create a learning and feedback loop which promotes high-speed innovation at a global level. Thus, with AI, the technology enterprises have doesn’t matter as much since it is widely available. Instead, the competitive edge comes from the training data set they have to ‘feed’ their AI. 

The importance of collecting digital exhaust

ML processes work by ‘reading’ large quantities of data, noticing patterns and learning, based on the data and instructions that they are provided with to better perform a task. The more data a machine receives, the faster it learns and the higher its accuracy rate will be. Exposing AI to large quantities of diverse data also decreases the risk of making biased decisions. Data is the driving force behind ML and, implicitly, behind AI. More so than the AI technology, it is the wealth of data a company has available to use with their AI that is going to be their asset. 

Collecting data – as much, as relevant and as early as possible – about business practices should be a key consideration for businesses looking not only to keep up with their competitors but to survive. If two companies with a similar funding, set of customers, budget and opportunities are only differentiated by the fact that one of them has started collecting information about how it does business sooner, that company is always going to be one step ahead. Every piece of digital exhaust that employees are emitting right now, is what companies will need most in the future. With machine learning, if your company doesn’t have the training data because you missed the opportunity to start collecting it early, you will quickly fall behind.

Upskilling the Workforce

As AI becomes a definite part of business strategies, there is a second consideration companies should have in mind. AI is more efficient than humans at looking through large sets of data and doing menial tasks, so many of these low value-added functions are going to be fully automated in the future. As such, some employees will find their workloads diminished. Companies are going to have to revaluate their division of labour. AI typically leads to increased productivity at the expense of replacing low value-added work, such as manual data entry, but not complex human labour. Training people to do high value-added work which can’t be automated can ensure a smoother implementation of AI technology into the business. 

This is where we can also think of AI as a co-pilot. Unlike autopilot, which is used to automate easy tasks and low value-added work, co-pilot is used in a complementary manner, offering guidance on how to better carry out a task, rather than automatically doing the task for the person. In this way, co-pilot AI can be used to provide insights to help workers more efficiently carry out complex tasks that can’t be executed by a machine. 

Starting early is the only way of keeping up. Implementing AI and solidifying it as a clear part of the business strategy, collecting as much data about how business is being conducted, and being ready to upskill the workforce to be able to work alongside AI and boost productivity should all be key considerations for businesses. Companies in the past have died out because they missed the automation train, an issue that could have been resolved with the right resources and technology.