Utilities have been consistently investing in addressing two key challenges: Improving revenue and boosting productivity. Both are severely impacted when customer data, meter reading and billing processes go awry. Manually identifying, investigating and resolving exceptions in these areas of business operations can be slow, tedious and expensive. They can lead to serious legal and regulatory impact, affect cash flow and create unpleasant headline news. For example, Npower - Energy ¹ in the UK was in the spotlight for £26m in fines it had to pay for sending out late and inaccurate bills. Scottish Power's ² inadequate standards for call handling, complaint resolution and billing resulted in over 1 million complaints in the space of 2 years, adding to unnecessary operational costs. Horizon Power ³ employees in Australia have had to face the wrath of customers in public due to under billing that ultimately, when corrected, led to a bill shock causing stress. Earlier this year in July, the Los Angeles Department of Water and Power (LADWP) ⁴ paid $44 million in settlement for a class-action lawsuit that accused the utility of over-billing. Artificial Intelligence (AI) has now come to the rescue of utilities with an effective answer to this phenomenon that has troubled the industry for decades.

Typically, teams of managers and executives wade through different applications to determine exceptions and anomalies, conduct a root cause analysis (RCA) and create resolutions. Each problem can take between one to four hours to resolve. However, Athena – an AI-based solution on Wipro’s HOLMES^TM platform for Oracle Utilities users – can reduce this to less than 2 minutes.

Athena sits on top of the Oracle Utilities Billing Component and learns everything that an exception manager does. It then begins to mimic the processes and actions of the managers. It simultaneously uses Machine Learning and, eventually, over a period of time resolves exceptions on the fly without human interaction.

Athena works by validating every solution it comes up with (“I am 80% certain of this solution. Do you agree?”). As it learns, it increasingly avoids going through process steps and reaches near perfection. Athena has shown that exception resolution process efficiency can be increased by a remarkable 95%.

The end result is that customer issues are resolved faster and more accurately, improving CSAT; there is a reduction in disconnections, overspends and late fee collections; cash flows improve; there is a notable reduction in human dependencies, leaving executives with more time to fruitfully interact with customers; and regulatory penalties, legal costs and public embarrassment are avoided.

There is an abundance of opportunities to use AI and Machine Learning across the utilities value chain that has traditionally been manually driven. Utilities are examining the use of these technologies to improve customer engagement, impact infrastructure management and reduce outages. With Athena, utilities can quickly harvest the upside of AI for exception management, resulting in revenue and productivity improvements.

References

1. https://www.theguardian.com/business/2015/dec/18/npower-ordered-to-pay-26m-for-customer-service-failures
2. https://www.theguardian.com/business/2016/apr/26/scottish-power-fined-18m-pounds-for-poor-customer-service-ofgem
3. https://www.abc.net.au/news/2016-05-16/calls-for-wa-horizon-power-overcharging-investigated-broome/7417390
4. https://www.scpr.org/news/2015/08/17/53836/ladwp-overcharging-settlement-means-44-million-in/