[ad_1]
It is hard to believe a bullet train is piloted by a human in the cockpit. Would you let this person take over the train at 300 miles/hour? In such a technologically advanced world, critical industrial processes such as inspections are inefficient, costly, and wasteful.
Countries are aware of their inefficiencies; according to a study by the International Monetary Fund, countries waste about one-third of their infrastructure spending due to inefficiencies. As digital technologies are integrated into business processes and workflows, McKinsey estimates that could lead to a two percent productivity boost, on average, per year over the next decade.
Efficiencies and Effectiveness
Industrial transformation takes many shapes and forms; traditional approaches include BPR, Six Sigma, and digitization of workflows. Today the game has changed. Three megatrends drive the transformation of entire jobs in the workplace. Large infrastructure is set to gain the most for their businesses on the path to autonomy through:
- The rise of A.I.
- The increased interconnectivity of machines and sensors
- The digitization of data
Early adopters of these topics will have a major competitive advantage, both for customer differentiation and operational efficiency.
It is essential to understand that automation is fundamentally different from autonomy. While automation of workflows provides higher quality and often faster workflows, autonomy provides a feedback loop to respond to fuzziness in the workflow.
While the drive was towards taking variability out of the process in the past, if done right, the new processes can live with fuzzy information while executing a job.
Take the speed control in a car as an example. In its simplest form, it automates the job for the driver and keeps the speed constant. Once you add more interconnected sensors to the system, you are creating a more sophisticated autopilot.
As you increase the number of sensors and the related compute power to make fast decisions, you are moving to autonomy – the system will automatically respond to unforeseen obstacles or hazards.
These responses to situations are even more effective if the human does not need to write all the rules but leave it to the machine to learn patterns. The current state of A.I. is particularly good at pattern recognition and then triggering a response.
As a result, A.I. enabled systems will have the greatest impact on workflows that are currently routine, predictable, and repetitive.
Large and vast industrial processes lend themselves very well to be one of the first beneficiaries of A.I. We can constrain the operational variability sufficiently to test and optimize the autonomy. Later, the operating environment can be generalized for greater benefit.
Finally, human-machine interfaces will continue to play a big role; only the nature of work for the human will change, moving from operating the process to supervising the process.
To move towards autonomy in industrial processes, one has to dissect and look at what can be fully automated, what elements will require human supervision, and which areas will remain manual. Once this clarity is established, one can set a path to autonomy. Starting with hardware, connectivity, compute and data which exist today and gradually inserting additional elements overtime to drive efficiencies and higher quality processes.
Real-world examples
Let’s go through a hypothetical example – assessing insurance claims after a large scale fire such as what transpired in Malibu, California.
Today, the community and insurance companies need anywhere between 6-9 months and thousands of experts and volunteers to get a precise overview of the impact. Many people are put in harm’s way and families are left with little or no information for many weeks. Thanks to the technology available today, this can now be done within weeks, sometimes even days.
Transforming this inspection process starts with getting the right equipment to be utilized in the field for inspections, i.e.network connected robots and drones equipped with computer vision and other sensors. Humans would manually set the flight area and clear the airspace. Utilizing a custom trained set of AI to collect, understand, and analyze the datasets obtained from the wreckage, the robots and drones would find their way autonomously to collect and transmit the data back to a human to derive insights for claims and recovery logistics.
This is real and ready for business today. Insurance companies can have certainty for payouts; councils can start the recovery effort with utilities; families know when they can start to rebuild their lives.
(Photo by John Cameron on Unsplash)
[ad_2]
Source link