Artificial Intelligence (AI) is an extremely hot topic these days, but just stating that a solution or application has “AI” in it has become a bit too popular. The key questions or topics should be “What exactly is AI? How is it being used in your system now and what are your future plans?”.
Internet searches on “AI” deliver a multitude of results on the topic to help you learn and understand the latest state of AI, the ongoing research, as well as many interesting use cases. This blog will discuss some of the use cases where AI can make an impact for CPSs, as well as end users utilizing next-generation applications and services.
Before we look at some use cases, we need to acknowledge that there is a massive shortage of AI experts around the world. The needs far outpace the education systems’ ability to generate highly educated individuals who not just use AI systems, but build and evolve them commercially. Google leads these numbers with over 1,400 experts on their roster; IBM and Microsoft follow with close to 1,000. This provides us with some clues as to the reality of AI:
- It is above three companies who will continue to make large leaps in AI evolution, and as a result their products and services will advance ahead of other companies that simply state that they are AI “enabled.”
- Even more important for service providers is the fact that if they are to leverage AI in their applications and solutions, they will need to incorporate solutions from one of these three camps to advance their cause.
Let’s first look at a relatively simple use case for CSP customers – online marketing. Are you tired of getting an offer from your CSP suggesting a product you already have? Or worse – a product that is either older than you own or one that is not applicable to your situation? For example, why would you purchase an iPhone 8 if you already own iPhone X? Or why are you getting an offer for a graduation gift for a high school student, if your children are already in University? It seems simple, but when you have millions of customers, perhaps it is not as easy as we might think. Machine learning (a subset of AI) keeps an eye on the increasing amounts of data that are collected and starts to deduce some important pieces of information from daily online interactions in order to present offers that make sense.
I know some of us are freaked out about privacy and “big brother” looking at our stuff, but if this is done carefully the outcome is targeted marketing on things that I want, I need, and I’m getting exactly when I’m about to look for it. If you ever do a Google search for a drill and then find it “shocking” that Facebook or Instagram is serving you ads of those items, that’s ML already happening and there is nothing you can do about it! I believe there is a lot of room for improvement in this area to better re-target the prospective buyer, by delivering on improved AI analytic results and machine learning. For instance, I do not want to see ads for stuff I just purchased, but rather more complementary items. Another example is seeing advertisements for a Ford truck, if I drive a BMW. So Mr. (or Ms.) CSP, can we invest a little bit of effort into delivering a more focused, re-targeted marketing strategy? I’m sure your customers will appreciate it and they might actually purchase something they do not have and perhaps even something they didn’t know was available.
Ok let’s switch gears into customer service. And BTW, I’m not throwing these use cases to poke CSPs in the eye. This is an open discussion to share ideas and thoughts, which I’m sure everyone can relate to and to provide some insight on how we can jointly solve these challenges.
So back to customer service. What happens then you want to upgrade your mobile phone, change your plan, or perhaps you are going on vacation to Jamaica and want to get that travel pack enabled to send those cute selfies to the your friends and family at home who are not with you in the sun? You might try to self-service your needs on the good old website, but if you have a question or cannot accomplish your mission or perhaps you are driving, you are going to hit that 1-800-help-me-out-mr-CSP. And what happens next?
You hear, “Hello and welcome to (insert your service provider name here). For your account balance press 1, for sales press 2, for business press 3, for store near you press 4,” blah blah blah. IVR should have been dead with the invention of smart phones. I honestly have no idea why these systems are still in operation. If you want to see something really cool, ask Dialogic about Dialogic Visual IVR which allows for an easier and more streamlined process to get the information you want by removing the need to press buttons on the phone keypad like it was 1987. But this is still not where AI comes to rescue. What happens in most cases in customer support when you reach that automated service where an “event” needs to happen, e.g., something needs to be updated, or purchased or added? You get connected to a human – someone in a call centre that is ready to help you. “Hello, can I get your mobile number, address, date of birth, shoe size etc.?” Not only am I calling from my home number or my mobile phone that YOU SOLD ME, but I am sure the call center agent has the ability to reference the number I am using. And we can’t ignore the fact that I just typed my account number 3 seconds before the call centre agent asked me for it again – lovely.
OK, rant over.So where are we going? It would be cool to get to a spot where two things happen:
- A system with which I could talk (or type like I text with my kids) in a free-form manner to discuss my needs with an automated system – a system that would understand what I’m asking, follow up with extremely intelligent next steps, and understand everything about me, my devices, my history, my usage, my family, and my plans without constantly saying “ah ok - let me put you on hold for a few minutes while I reboot my screen.”
- A system that based on our conversation would be able to actually finalize our “transaction” – update my plan, give me my roaming package, or get me that new phone I’m eligible for.
The speech recognition and AI technology to do this is definitely in place, is secure enough to recognize your voice plus identify all fraud cases (banks are using it now), and able to make the transactions happen. I predict that more CSPs will start implementing solutions like that in 2019 and 2020. Some have already started.
There is a gap, however, that needs to be bridged between CSP applications and the three leaders in AI – Google, IBM, and Microsoft. All three use SaaS models to allow anyone to pay fractions of a dollar to identify the intent of a conversation, carry the dialog in an intelligent manner, and make sure it all happens in a human-like fashion. If you are an application vendor where a user might need AI once or twice per month, that is very inexpensive. For tier1 CSPs with millions of users making thousands of requests per day, that quickly adds up, especially in the current extremely competitive CSP environment. So how does a CSP leverage top AI engines and make it cost effective for their operations? Enter DialogicOne. The ability to abstract interaction with AI engines for specific use cases gives DialogicONE a leg up in those specific use cases. DialogicONE can not only help CSPs save money, but can also further provide a technology model in which the servers allow users to interact with CSPs via mobile applications, smart speakers, and their website in a consistent fashion.
If you are a CSP and are thinking about leveraging an AI engine to give your users an exceptional experience, differentiate yourself from the competition, save money on call centres, and improve your target marketing, then let’s talk. DialogicOne is utilized in building solutions for operators around the globe and we can certainly discuss how this could help improve your existing business models.