Why Natural Language Technology Will Let You Down

[24]7.ai | January 30, 2018

We live in a world dominated by technology, devices, and online interactions. Siri, Alexa, and Cortana are becoming staples in households around the world, offering us the capability to interact with technology and virtual assistants the same way we'd interact with friends or family.

As we become more accustomed to talking with these devices the way we'd talk to another human, it's reshaping our expectations of all interactions with technology. We now expect the ability to talk or type conversationally when interacting with companies on their digital channels, and to have our issues easily understood and resolved quickly.

This can present a problem for organizations who aren't using the right technology in their customer experiences.

What Natural Language Can't Do

The interactions were used to today are possible because of natural language applications that allow technology to understand and derive meaning from what we say or type, and offer accurate, insightful responses. This technology is a core part of most customer experience strategies, but to work effectively and deliver seamless experiences in line with modern-day expectations, it needs to be paired with predictive technology that can determine actionable intent.

Our research has determined that natural language technology on its own has a 40 percent failure rate in identifying actionable intent.

To give you a better understanding, let's look at a common example of how natural language technology on its own can let you down.

Scenario: A customer is attempting to pay for an online purchase when her credit card is declined.

Thinking she may be over her credit card limit, she logs into her online banking account and sees that she is not. She calls her credit card provider and reaches a natural language IVR. The IVR asks, "How may I help you today?" She responds, "My card is not working."

Based on this statement, a natural language solution would narrow down the range of probable reasons to:

  1. The card has been blocked due to suspected fraud.
  2. The customer reached her credit limit.
  3. The customer's payment is past due.

The IVR would then offer the option of speaking with a fraud agent or a billing agent as the next step. These might be the correct options, but wouldn't it be better if the credit card provider knew about any preceding activity that the customer took? In this example, knowing the customer visited her online account would help to determine the most optimal path to resolution for her issue.

This is where predictive technology comes into play. The predictive platform would know who the customer was and what she was doing online before calling in. Combining that knowledge with her account profile information, it would determine the problem was an overdue bill, and offer the correct solution. The IVR would give the customer a scripted explanation about why her card was declined and immediately provide the option to pay the balance owing.

Compare this to the scenario offered by natural language alone, which would require the customer to transfer to a call agent where she might have to restart her journey, and it's clear to see that combining natural language with predictive technology is essential to providing faster journeys with better resolution rates.

Start exceeding your customer's expectations. Learn more about how we can help you combine natural language with predictive technologies to provide optimal customer journeys.