I have long been deeply passionate about artificial intelligence (AI), having dedicated my professional life to data science and researching multiple aspects of AI, including Natural Language Processing (NLP), time series analytics, and how to make insights actionable.
Applying AI to attaining the business objectives of service providers is fascinating. Whether for marketing analytics, social network analytics (SNA), or IT operations analytics (ITOA), what AI can do for service providers is nothing short of remarkable. For example, by leveraging AI paradigm-shifting technology, we can detect and even predict anomalies to improve so many aspects of the business.
Turning the day-to-day on its head
The thing is, even with automation, AI is definitely not just about reducing costs and accelerating performance. Though this seems to be the focus of AI talk these days.
Through my extensive and long-standing affair with AI, both in the context of the communications and media service providers industry, and as a university instructor – I can firmly assert that the beauty and elegance of AI is about shifting our mindset, doing things differently, and turning traditional work processes on their head . . . for the better, of course! Whenever I speak with service providers about how AI can help them overcome their challenges, they start out by asking for help with costs, speed, and quality. But, once we get started, they quickly see that AI and data science can bring them much more.
Got knowledge? Then enrich it (with AI)!
For example, we were engaged with a service provider who wanted to take the manual work out of sending tickets into automation flows. To enable this automation, we introduced the idea of using NLP to classify tickets, specifically those that have a high recurrence, into pre-defined categories.
After capturing a few quick wins from the automation of recurring tickets, the service provider came back with additional requests to automate more complex tickets that required a lot of training and a great deal of manual work on behalf of their service reps. Just training employees on how to handle these complex tickets entailed 3x the cost of investing in AI. But it’s not just about the hourly cost of manual labor vs. AI, nor just about saving time. It’s not even about the value of being able to make a high quality resolution sustainable.
And here is where the elegant part comes into play – AI is about aggregating, enriching, and protecting knowledge. When the knowledge of how to handle complex tickets is maintained, leveraged, and enhanced by the machine, the risk of losing that knowledge is mitigated. Let’s get industry-specific When we’re talking about complex tickets, developing an algorithm that classifies such tickets into automation flows, means having to deal with many challenges during the NLP training phase.
To illustrate, service reps usually use a certain shorthand when writing up the tickets, which the algorithm needs to learn how to decipher. Moreover, reps in non-English speaking countries often transliterate – that is, they use the Latin alphabet to spell out words in their native tongue. Needless to say, deciphering their notes is no simple task.
There is also the issue of being industry specific. Namely, the service provider’s ‘dictionary’ includes a long list of terms and phrases that are unique to our context. Several names can be used to refer to the same application. And, when we say ‘number,’ we need know whether we’re referring to the customer’s phone number . . . or ID number . . . or the customer indicator.
Get the picture? It’s complicated!
This is the exciting side of things – overcoming the complications with very advanced data science.
The power of continuous business insights
We have seen that AI is perfect for resolving tickets automatically, both simple recurring tickets, as well as infrequent and complex one.
Though, there is another side to the elegance of AI that I am also very passionate about – and this is the fact that AI can be empowering. AI can revolutionize how CSPs actually do business, beyond the IT operational side of things. For example, we support our customers in their executive decision making by using AI to extract and deliver insights from data that would otherwise not be available.
Decision makers are typically flooded with data. It’s impossible to know what to focus on. There’s no way to go through it on your own and pick that strategic needle out of the data haystack.
The analysis of data by data analysts is a long process that requires high levels of expertise. Therefore, seeking out business insights is traditionally done on a monthly or even quarterly basis. Needless to say, this kind of turnaround does not align and cannot keep up with the speed of modern business.
But, with AI, the insight selection process is automated, intelligent, and fast! The algorithms know what to look for, how to score what they find, and then to push those insights to the right decision maker for action.
For example, with AI, important anomalies such as trends regarding top-ups, timing of payments activities, SLAs regarding shipments, and much more can be detected.
There’s no doubt that AI-powered tracking and alerting brings a significant and measurable impact to the business.
So, if you’re also passionate about what AI can do for your business, I invite you to reach out to me so we can discuss how you, too, can harness the power of AI.
About the author: Dr. Gilli Shama is Chief Data Scientist, Amdocs Data Intelligence