Will predictive AI finally solve the multi-billion-dollar downtime problem?

Downtime for today’s large, complex businesses means more than a simple inconvenience. The cost of interruptions, especially when workers are prevented from completing tasks due to out-of-service infrastructure, can be huge. A Gartner study called “The Cost of Downtime” suggests that a large company may actually lose as much as $540,000 per hour from a preventable technical failure.

For service companies, that means providing equipment upkeep and repairs in a timely and effective way to reduce unplanned maintenance. The challenge is that a technician on a call may not always possess sufficient knowledge, or have the right tools and parts on hand, to expedite unplanned service. Service companies simply don’t have the inventory to enable every technician to carry every part that might fail. So what could have been a simple repair taking minutes becomes a dragged out affair? That’s downtime in a nutshell – and globally it’s a $647 billion a year problem.

The problem is that without a holistic assessment of equipment wear and reliability, scheduled maintenance may be out of whack with reality. Why is this? Up until recently, service schedules were based on intuition, experience, and happenstance. But these common strategies, like run-to-failure, scheduled maintenance, and reliability-centered maintenance (RCM), aren’t much more than lucky guesses. Service companies are now increasingly looking to adopt predictive analytics to eliminate time wasted and maximize profits attained.

Not surprisingly, a new generation of predictive analytics companies are tackling this multi-billion-dollar problem. NY-based Aquant.io, for example, is bringing a new, comprehensive approach to the downtime problem. Using predictive AI and machine learning, they’ve developed algorithms that anticipate where and when service is needed and which parts the technician requires on hand. They call the strategy “selling uptime.” The company is the brainchild of Shahar Chen and Assaf Melochna, who recognized the potential of predictive AI after 15 years of business and technical expertise in enterprise software, and consulting for Fortune 500 companies on field service management. The algorithm research and development was led by a machine learning expert who applied groundwork from the neuroscience world.

Here’s are five ways predictive AI is being implemented by service companies to save money and be more efficient:

Maximizing productivity of equipment and technicians

Even expert service people aren’t clairvoyant. They may not know exactly which parts and tools to carry, and may not be able to anticipate precisely when a fault may occur. Predictive AI fills the gaps by analyzing statistical data at a scale that’s impossible for human analysts. By data mining a wide range of relevant variables that can lead to failure, maintenance schedules can then be developed that are efficient, and save time and money.

In fact, this strategy has already been applied in other industries with great results. California’s Imaginea has been using predictive analytics in the hospital industry to overcome staffing, overtime and burnout challenges by looking at resource availability, staff schedules, operating hours, appointment history and hundreds of other factors.

Reducing the need for repeat service visits

With predictive AI, a service company can cut down on multiple service visits. In fact, some studies suggest that one out of four on-site visits is a repeat visit. At an average of $200 a call, that’s a lot of money being spent with no positive results. Predictive AI can help by:

• Predicting the required parts and skills

• Solving by the most cost-effective way

• Optimizing parts management for each service vehicle and location.

Analyzing fault histories and providing service action recommendations

We’ve already mentioned this point above, but it’s so important that it deserves its own section. The power of this benefit isn’t to be downplayed. Using predictive AI software, your business is now armed with AI and machine learning. The result is the ability to accurately provide recommendations, including:

• Predicting which faults may occur next.

• Providing time-to-failure predictions.


• Determining what the most cost-effective solution is to a client’s current problem.

• Providing estimates for parts, skills and job duration (triage assistance).

• Providing a step-by-step troubleshooting process.

Make better business decisions

One of the strongest use cases for predictive AI in any industry is the ability to make better, data-driven business decisions. Predictive analytics is already being applied to optimize production and key processes, and improve customer service. Predictive AI will comb through all available data and provide you with actionable insights for your business. This can be delivered to executives and managers, allowing the business to run smoother and in the right direction. In other words, toward higher profits.

Gartner predicts that “by 2018, more than half of large organizations globally will compete using advanced analytics and proprietary algorithms, causing the disruption of entire industries.”

Optimize pricing

One of the biggest areas in which service companies lose profit is when they fail to optimize pricing for service and parts. Service companies can actually use real-time pricing optimization like Amazon. By changing pricing based on parts availability, service demand and the likeliness to achieve your target EHR, you’ll maintain profitable margins.

The only way this can be done is through predictive AI. The combination of AI and machine learning will calculate all available data to help you more accurately know how much to charge each unique customer.

Conclusion

The headlines have been filled lately with well-publicized technical failures. Last year’s Delta Airlines power outage, for instance, resulted in disruptions for thousands of travelers. In Delta’s case, the problem was triggered by a failed switchgear. More efficient maintenance may have prevented six hours of downtime for one America’s busiest airlines and saved the company millions of dollars.

Given the potential to prevent these sorts of massive technical failures, and save companies money lost from preventable downtime, predictive AI seems likely to be the industry standard very soon.