A rapidly growing number of people believe that AI is going to transform at least 25% of the workforce in the next decade. That’s according to McKinsey’s research. The USA is going to hire skilled staff to populate these roles and Europe is going to retrain current staff. Whoever gets this right, it will lead to a major disruption in business operations and we have seen this first hand in most of our clients.
AI is coming but it’s a lot slower than you might think. Most of the stunning bleeding-edge capabilities which we see demonstrated today produce very risky outcomes or sit in novel use cases which have questionable business value.
However, the mainstream AI technology has no infrastructure on which it can be effectively implemented so that we can guarantee ROI. As Microsoft and Facebook have found Chatbots are not necessarily good value and other great innovations are playing games well but fail the business case test.
The biggest challenge is that businesses are not a fixed product in which we can build a constrained AI. They’re an ever-changing mashup of markets, customers, products, finance, risk, innovation etc and AI needs to scale, and the platform to scale AI is automation.
Typically we see automation technologies in the market will have a level of what we can arguably call AI. It’s logic, 2 + 2, if this then that, that is a form of AI. It’s a very low level and it’s not commonly known as AI but it does make automation intelligent. It enables them to make decisions, however you may find that there’s a number of vendors out there that will try and convince you that putting some type of the more advanced AI and plugging it in to that automation will enable you to achieve your goal but as you’ll see from this visual, there are some gaps in that.
The first thing is we don’t have all the technology today. We don’t have anywhere near all the AI technology to be able to replace a human being. Listening to the popular news about how machines are going to take my job, put simply they can’t.
But we need people to be able to interact with machines so we need that human-machine interface and you need to be able to bring that experience and the human reasoning over and above AI to be able to deal with this. AI is also driven by different goals. They’re not always exact, not always specific outcomes that you’re looking for. Sometimes it’s a goal or objective which may or may not actually ever be achieved but you’re constantly seeking that goal and that’s where we really get into that higher level probabilistic technology and we’re a little bit intolerant of that. We don’t trust the technology. We need to be able to intervene, we need to be able to sit in the driver’s seat ready to take over, but we can trust the machines to do a little bit.
To bring that together you need to orchestrate, now orchestrate is not integration, not how a lot of vendors may tell you we just need integration, we just need to talk to third-party systems, we need to talk to IBM Watson, to talk to your chatbot, it’s not true. You need to orchestrate, orchestrate is an integration that is driven by decisions, not by action, it’s the integration equivalent of AI not the integration equivalent of automation.