Guest Blog: Debunking Five Artificial Intelligence Myths – Real Tales of AI in Customer Experience

This week on our Friends on Friday guest blog post my colleague, Vince Jeffs, examines the hot topic of artificial intelligence (AI) and it’s impact on the customer experience. I am reading articles every week about AI and find it’s impact fascinating. – Shep Hyken

If you’ve read or seen any media lately, you know that the Artificial Intelligence hype is off the charts. Recently, even 60 Minutes Overtime did a segment on robot research at Carnegie Mellon, further fueling the hysteria. The imagination runs wild as to what this all will bring to our future.

And sure, I too enjoy dreaming about AI as well. But the reality is we receive paychecks because our employers expect us to deliver commercial value.  How can we take this out of the realm of science fiction and into reality?

So rather than falling into a trance that AI will radically transform the world, let’s unpack what is commercially real in how this translates to delivering better customer experiences.

Below, I’ll debunk five myths and contrast with actual examples of AI in use.

Myth #1: AI can think and operate like humans today

Machine intelligence can’t match the human brain. Most experts agree that the “Turing Test” and other methods for gauging AI progress show we aren’t there quite yet.

In CX and CRM, however, we benefit from machine learning algorithms and robotics because software can access and traverse massive data sources and find patterns we’ll never see.

Subsequently, brands from telecommunication companies to retailers have seen impressive results. AI-powered product suggestions for consumers regularly improve response rates by 300 percent or more. I love when brands find helpful items that I never considered, such as an accessory with my new iPhone, or a shirt with the shorts I just carted. This not only drives increased revenue but also provides unexpected value to the customer.

Of course, none of this replaces human thinking. Instead, these technologies assist us and make our work and experiences more productive.

Myth #2: The rise of AI means machines will completely overtake the workforce

Successful companies rarely destroy jobs. Instead, like the laws of matter – they morph into new forms.  For 30 years, businesses have improved efficiency by streamlining redundant tasks to free employees to do more meaningful work. I expect brands to use technology in a similar fashion in the next 30. In many cases, AI technology works hand and hand with people to do work better, smarter and faster. As firms reengineer outdated processes, they transform them – removing some roles, but creating brand new ones.

AI can pinpoint waste and repetitious patterns, enabling managers to make solid defensible cases for new projects, tearing apart dysfunction and rebuilding it. For example, robotics software installed on desktops helps isolate and automate wasteful and repetitive tasks. Employees become productive and happier, focusing on providing a better customer experience.

Myth #3: I need to hire a gaggle of Data Scientists

Make no mistake – you need data experts and employees that get AI science, but you don’t need dozens of PhDs to accomplish your AI goals. Instead, concentrate on hiring and assembling a small team with the right mix of talent – those who understand machine learning technologies, applied statistics, your business, content management, and project management.

Make sure the team has clear objectives and milestones and give workers incentives to turn their ideas into profits. At a large bank, I’ve witnessed a slim four-person team build high performing propensity and learning models for all business lines.

Myth #4: I should save all my data because AI will need it

Data is a cost. Like old stuff in an attic, it just ends up occupying space and is useless until you pull it out and use it.

You need to store data that might one day pay off. But how do you know what data is gold and which can be set to the curb? Consider these tips:

• Use common sense, business intuition, and convenience when deciding which customer attributes to store.
• Keep purchase data (first-party data) for up to three years or longer for infrequent purchases. Summarize for quarterly and annual trend spotting and forecasting.
• Tag key digital interactions and keep this behavior data longer than preference and survey data, which goes stale.
• Overwrite old demographic profiles and modeled data (third-party data), which yields little value.

Myth #5: AI is some super new science that I’ll never understand and use

AI is a broad area and encompasses some steadfast, road-tested science, such as regression models, decision trees, and emerging techniques like deep learning.

If you aren’t applying any of these methods, start with simple and low-risk approaches. Use technologies that employ Bayesian algorithms to identify next best action based on response data, or find compatible products to offer based on collaborative filtering.

Once you achieve quick wins, work toward using models that calculate Customer Lifetime Value and Churn Propensity.

Final Thoughts:

Sure, Artificial Intelligence sounds like the stuff of tech fantasyland. To stay competitive, hunt for vendors and technology that can demonstrate real value to the business today. If you can cut through all the hype, you can find AI that turns customer data into actionable insights that return significant value for your customers.

Vince Jeffs is the Director of Strategy & Product Marketing for Pegasystems.

For more articles from Shep Hyken and his guest contributors go to customerserviceblog.com.

Read Shep’s latest Forbes Article: ‘National Returns Day’ Is A Lesson For All Businesses