An Interview with James Nies from Verint Systems & Muffy Pease from Salesforce on Artificial Intelligence & the New Customer Service Paradigm
(Keith Fiveson, MC for the NY PACE event at EmblemHealth in NYC on 5/18) Thank you for taking the time to participate in this “pre-event interview.” We wanted to get a jumpstart on the discussions we will be having at our upcoming event on the very wide range of topics we will attempt to cover relating to AI in the contact center, including Chatbots, and Robotic Process Automation.
(Muffy Pease, Healthcare Technology Specialist, Salesforce) Let me start by telling you what AI is, in Salesforce terms ... our solution (we call it "Einstein") enables our customers to use their data to make it easier to sell and provide great customer service. Here is a quote from our website:
"We live in a hyperconnected world where every digital interaction — from phone call to purchase to page view — adds to a never-ending onslaught of data. And with the advent of the Internet of Things (IoT), even inanimate objects like cars, refrigerators, and clothing generate additional data by themselves millions of times a day.
All this data can be used to increase sales, fine-tune marketing, and provide the immediate and personalized service today's customers want. But how can your business turn a bottomless ocean of data into the steady stream of insight needed to fulfill those expectations? Artificial intelligence is the answer."
How do I assess my organization's adaptability for AI and RPA (Robotics Process Automation), and what is the ROI for changing what we are doing today?
(Jim Nies, VP of Product Strategy for Verint’s Digital Engagement Solutions) RPA is actually really very easy to adopt compared to other solutions you’ve probably deployed in a customer service or back-office operation. A successful RPA project doesn’t require an alignment with a broader customer experience strategy or wrestling with technical compatibility. RPA solves a problem of efficiency, accuracy, and cost – I’ve yet to meet an organization not on board with that. It can leverage existing human interfaces, so there isn’t a big question technology readiness or direction.
I’m not sure it makes sense to think about AI distinctly in terms of adaptability or ROI. AI is going to increasingly be part of solutions you put in place. Better self-service capabilities will provide ROI and those are likely to contain AI, but that’s a self-service question, not an AI question.
How do you define Artificial Intelligence and Machine Learning?
(Jim Nies) The most concise summary I can provide is: AI is a concept and machine learning is
Many people are familiar with the Turing test, which was devised to answer the question, “Can machines think?” Notably, Alan Turing asked that question before the first integrated circuit or silicon chip existed. It’s a stretch to say the Turing test is AI in a nutshell, but it’s an adequate starting point.
Most people also understand their lives are impacted by predictive algorithms – which are probably the classic example of Machine Learning. Credit scores, for example, or the particulars in the offers that clutter our real and electronic mailboxes have predictive algorithms behind them. The more sophisticated implementations of these algorithms can improve themselves over time. Strictly speaking, that “self-tuning” is not required to meet the definition of machine learning, but increasingly, I think it’s expected when people hear the term “machine learning.”
Machine learning algorithms are also used to recognize patterns – including for image - or make predictions in ways that don’t simulate human thinking. If I want to create an “artificially Intelligent” system—a system that simulates human thinking—it’s likely that, among other things, I’ll need to use machine learning algorithms.
(Muffy Pease) Artificial Intelligence (AI) is the concept of having machines “think like humans” — in other words, perform tasks like reasoning, planning, learning, and understanding language. While no one is expecting parity with human intelligence today or in the near future, AI has big implications in how we live our lives. The brains behind artificial intelligence is a technology called machine learning, which is designed to make our jobs easier and more productive.
What does artificial intelligence mean to you?
(Jim Nies) To me, it means a more natural and efficient way to get a computer to do what I need it do. This is partly an “interface” problem, using AI to enable me to get a command into the computer in a more natural or effortless way. It’s also a “feature” problem, using AI to make the feature “smarter,” such as winning at chess or selecting an ideal solution for example. I think we’ll notice more early progress on the interface side, but there’s already progress on both.
(Muffy Pease) One example of the impact of AI is Natural Language Processing. NLP is AI that recognizes language and its many usage and grammar rules by finding patterns within large datasets. One application of NLP that’s gaining traction is sentiment analysis within social media. Computers use algorithms to look for patterns in user posts across Twitter, Facebook, or other social networks to understand how customers feel about a specific brand or product.
What are some specific examples of AI technologies?
(Muffy Pease) Three examples you might be using today are Siri (from Apple), Alexa (from Amazon) and Cortina (from Microsoft) .... another great example is Amazon or Netflix engines that suggest other products you might be interested in, or similar movies to those you've watched and liked.
Is it a problem if AI can predict what we're going to do?
(Jim Nies) No, not fundamentally. It already does and it makes our lives better.
It’s a problem if a prediction is used unfairly against someone, or if people are misled somehow. This is already a problem society wrestles with and regulates. I’m not a politician, but I think people need to achieve a degree of literacy on this kind of technology so they can make decisions that serve their interests. This includes decisions they make as individuals, as well as
decisions we make as citizens in a civil society that sets laws and regulations. It’s no different than the need for people to have a degree of financial literacy so they can manage their income, debts, and insurance.
What is the potential use case for AI in customer service?
(Muffy Pease) In sales, three potential use cases are:
Predictive scoring — When Einstein gives you a score, it will also give you insight into how it was arrived at. For example, predictive lead scoring gives each sales lead a score representing the likelihood it will convert into an opportunity. You also get the reasons behind the score — for instance the lead source, the industry, or some other factor is an especially strong indicator that a lead will or won’t convert.
Forecasting — The predictive capabilities of AI aren’t limited to scoring; they can also be used to predict the future value of something, like a stock portfolio or a real estate investment. If you’re a sales manager, AI can predict your quarterly bookings and let you know ahead of time whether or not your team is on track to meet its quota.
Recommendations — Anyone who shops online knows that AI makes suggestions for retail purchases, but it can also make smart recommendations for any other product or service category from business software to tax consulting to cargo containers. And AI can also recommend things other than products — for instance, which white paper you should email a prospect in order to optimize your chance to close a deal.
In Service with AI, an organization can provide proactive service by anticipating cases and resolving issues before they become problems... Or It can automatically classify cases and intelligently route them to the right service agent.
Are Intelligent Assistants the only application of AI for customer care, or are there other ways that AI technologies can impact the contact center?
(Muffy Pease) Absolutely not ... AI can help in many areas, from voice recognition, to automated chat bots, to helping the agent solve problems faster.
What types of customer service interactions represent the best opportunities for automated interactions?
(Muffy Pease) AI is already transforming your customers’ expectations. Think of the consumer who lives by Uber, Google, and Amazon. If he walks into a department store to buy a suit, what does it take to provide him the same level of service he’s grown accustomed to?
Retailers should know who he is because he bought something online. They should know his size and preferences based on his purchase history. And they should be able to suggest the perfect pair of shoes to go with whichever suit he chooses.
The same principle applies across every type of business. Customers know you have their data. They know everything you can do with it. And they expect you to use it to provide fast, smart, personalized engagement across every interaction.
Do we need a data police, like AI?
(Jim Nies) Yes we need it, and I’d argue that we have it now. We have laws and regulations that protect how an individual’s information can be collected, stored, and used—especially financial data, but not just financial data. Insurance rating factors are regulated, for example. I don’t know if we have it right, but I’m cautiously optimistic that functioning democracies will figure out the right balance and then get the regulations dialed in. The degree of variance worldwide, even across societies that otherwise have similar cultures and laws—such as EU data laws versus US laws—suggests we’re far from consensus. It’s worth mentioning education again: We need everyone to have a degree of literacy on this technology.
AI is and could be powerful. Should we have a campaign to stop the killer robots?
(Jim Nies) Do you mean a robot that kills errors and inefficiency? I know more about those robots and I’m ok with them. Or, robots that find and kill cancer cells? I know a lot less about them. I think we probably want them also, and we can regulate them like any other medical device.
Do you think we will see a tipping point where call centers won’t be necessary anymore?
(Muffy Pease) I don't believe Call Centers will ever go away. Agents will always be needed to deal with the complex issues.
Click here to learn more about the PACE NY Metro Chapter event on 5/18 at EmblemHealth in NYC entitled: Artificial Intelligence (AI) and the New Customer Service Paradigm