By Jim Morrish, Founding Partner, Transforma Insights.
At Transforma Insights we have recently invested a lot of effort developing a comprehensive set of Artificial Intelligence (AI) forecasts. The approach that we took was bottom-up, seeking to identify where, and why, AI is deployed. In this article we discuss some of the high level results, in particular examining the symbiotic relationship that exists between AI and the Internet of Things (IoT).
Firstly, it’s worth clarifying that our AI forecasts aren’t like many other AI forecasts out there. We took the approach of breaking down the market for AI into a comprehensive set of Use Cases, which range from things like Complex Image Processing and Natural Language Processing through to Risk Analysis and Threat Detection. The list extends to some 42 Use Cases which we think together represent the opportunity for AI deployed in support of enterprise endeavours (although we did omit AI in applications deployed on PCs, tablets, and handsets, other than in the case that those applications are extensions of enterprise processes).
Also, we didn’t forecast a dollar market size, since it can often be unclear what those dollar figures actually mean: are they spend by end users, spend on licence fees, spend on services that include AI capabilities, investment in services, or investment in research? Or a bit of all of these things? Rather, we forecast the Instances of AI based on Use Case and where the Instance is deployed. So, for example, we would record an Instance of complex image recognition on an AI-enabled CCTV camera, and so on.
This analysis tells us that the vast majority of AI instances will be accounted for by deployments on IoT devices. In fact, over 95% of Instances of AI will be deployed on IoT devices consistently over the period of our forecasts from 2020-2030. Edge instances of AI grow faster than IoT, and the proportion of AI instances that is deployed on Edge infrastructure increases over the period from around 0.4% of Instances to around 1.3%. Unsurprisingly, given the more developed stage of these markets, the share of AI Instances deployed in the cloud and on PCs, tablets, and handsets (where we do count AI instances that are extensions of enterprise operational processes) falls over the forecast period.
All in all, we forecast that there are about 2 billion instances of AI Use Cases live today, growing to over 20 billion by the end of the decade. This growth is illustrated in the chart below.
It’s interesting to look how the breakdown of these Use Case Instances changes over the forecast period (and as illustrated in the chart below). Currently, Natural Language Processing, Chatbots & Digital Assistance, Image Processing & Analysis, and Activity Recognition dominate, accounting for about 80% of the total count of AI Instances worldwide. These Use Cases will remain amongst the most significant Use Cases in 2030, joined by Customer Behaviour Analysis. More importantly though this newly augmented set of top applications will only account for 66% of AI deployments in 2030, so the long tail of AI will get longer and more important over the next decade.
Clearly the AI market is already significant, and it’s growing quickly. With these new forecasts we aim to support a discussion about how that market will develop over time: identifying where AI is deployed and why. Analysis of commercial opportunities associated with AI can flow from these forecasts with the overall profile and scale of the opportunity depending, of course, on who you are: opportunities will be different for different vendors depending on how they engage with AI markets.
This UrIoTNews article is syndicated fromIoTBusinessNews