Getting started on the right foot when it comes to AI can be daunting. Particularly, when we see some confronting statistics:
· IDC found for one in four companies, 50% of their AI projects will fail
· Gartner estimated that 85% of AI projects will fail to deliver
· VentureBeat AI reports 87% of data science projects never make it into production
In the case of the IDC survey, respondents stated that their AI projects failed due to a lack of skilled staff and unrealistic expectations. The second factor, ‘unrealistic expectations’, can arise when there is a misperception of what AI is, what it can do for a company and how they should implement it.
At Canvass, we believe AI initiatives should be approached as a journey rather than an endpoint. A good analogy of how we should approach AI is by comparing it to the advancement of how we drive cars – using manual gear shift, cruise control and to, the ultimate in AI in motion, autonomous/self-driving cars.
In the industrial manufacturing setting, an AI journey can take on a similar trajectory. The ‘manual’ phase emulates a lot of how manufacturing processes are being operated today whereby processes are being manually controlled and they usually take a reactive approach to manage their equipment. As industrial plants increasingly put in sensors, embrace the Internet of Things and monitor performance in real-time, they are putting their technology stack together for the ‘cruise control’ phase.
Much like in Cruise Control, where you can set the speed you want to drive on the highway but then take back control when you want to exit, the next phase of AI adoption allows industrial manufacturers to automate repetitive, manual tasks but when a variable falls out of spec, human operators then take back control to make the necessary adjustments. There are many examples of this happening in industrial plants today, check out our blogs to learn more about live use cases of Industrial AI that are empowering the auto, food and ag, oil and gas, renewable energy and metals and mining industries to run their operations with ‘Cruise Control’.
When we get to the Autonomous Driving mode, the car is controlling the navigation, speed, observing for obstacles and anticipating upcoming traffic jams. Essentially, humans are in the passenger seat. An important distinction to make here is that navigation control, speed setting, looking for obstacles etc, are individual AI solutions that are amalgamated together to power an autonomous driving car. This is the same in the industrial sector, whereby there is no single AI solution, but rather a number of AI solutions integrated together to foster automation in the operations plant.
A lot of companies embarking on their AI journey want to skip to the ‘Autonomous Car’ mode. However, this is where unrealistic expectations may be set and ultimately not met. Industrial companies need to go through the manual phase of connecting their processes and collecting the data to set the baseline of your operations and understand how and where automation can improve yield, optimize efficiencies or reduce costs. The ‘Cruise Control’ phase is critical for building trust in AI, which will then foster the environment where you can scale out your AI implementation into adjacent processes, operations and eventually across multiple facilities and reach the vision of ‘Autonomous Driving’.
To learn more about how to kick start your AI journey, check out this video “How to implement AI in your operations.”
