Knowledge and AI Augmentation: The Essential Complement to Agent Training in the Modern Contact Center

State of CX

CX performance has dropped, or stayed stagnant at best, over the last couple of years, according to most industry surveys. The Forrester CX Index trend says it all. The percentage of companies offering “excellent customer service” (score of 85% to 100%) has been: 1% in 2016, 0% in 2017, and 0% in 2018.

The Problem

Forrester Consulting surveyed 5,000 consumers on our behalf about their biggest pain points in getting customer service. The top hurdle, by far, was lack and consistency of agent knowledge.

Contact centers have been trying to address this issue with training—including onboarding and training new agents and providing ongoing training to existing agents—to the tune of $10K to $40K per year, depending on the company and the industry. In fact, US companies spent $70.65 billion on employee training in 2016 alone. Yet agent knowledgeability, in the context of customer service, remains a big issue. Clearly, while training is important, it is not able to address agent performance when the customer is on the line. A SurveyMonkey survey of over 600 contact center agents worldwide proves it. Not surprisingly, the agent pain points had mirrored the consumer pain points, with finding the right answers to customer questions being the top issue.

The Causes

Despite continued investment in onboarding and training, why are agents struggling to answer customer questions? The reasons are manifold, but here are some key ones:

• The famous Moore’s law not only applies to compute power but also to the explosion of information and content, even when it is specific to a business.

• Agents are expected to answer increasingly complex questions from customers, who are likely to be irate since they would have already tried out self-service in vain. In some cases (e.g., BPOs), agents may be serving customers of multiple clients, which compounds the challenge!

• Demographics, science, and legacy tools conspire to make their job even more difficult:

• Most of today’s contact center agents are millennials and, increasingly, Gen Z. Training is not their cup of latte. If you think millennials have a short attention span (12 seconds), Gen Z have a “gnat” attention span at 8 seconds. While millennials may tolerate some formal training, 65% of Gen Z would rather just learn on the job.

• Then there is the “forgetting curve”: A concept pioneered by Hermann Ebbinghaus, a German psychologist, it refers to the rapid deterioration of human memory over time. This memory decay is also called “transience.” Ebbinghaus’ research showed that humans retain only 25% of information after two days of learning something. Research by the University of Waterloo says we retain only about 2%-3% a month later. And then there are Daniel Schacter’s seven sins of memory, which describe, apart from transience, six more ways memory is sabotaged—absentmindedness, blocking, misattribution, suggestibility, bias, and persistence.

Contact center agents, too, are not immune to the forgetting curve. It creates the need for endless refresher training and adds to associated costs and CX degradation since a proportion of agents is always in training.

• Millennials and Gen Z like to live “guided” lives—GPS for driving, and robot advisors to manage finances. They would rather “look it up” or search for answers than retain knowledge in their heads. Instead, much to their dismay, they find themselves using legacy knowledge base tools, which make them wade through tens or hundreds of irrelevant search hits, pore over documents, or hop across multiple applications to find answers when the customer is on the line.

• Facts and figures are not the same as analytical and problem-solving skills. Knowing a company’s products and policies isn’t the same as resolving issues or providing contextual advice to the customer, depending on need. Training does a good job with the former but agents (and customer self-service systems) need a solution for the latter, which can also complement the former.

 

The Solution

The answer lies in a modern knowledge management (KM) system, infused with AI, that can guide agents (and customers in the case of self-service) to the answer needle in a haystack of content and through customer conversations. Modern AI-infused KM also addresses all the other issues mentioned earlier: explosion of and constant changes to information, increasing query complexity, next-gen attention span, “all-gen” forgetting curve, “guided” lifestyle, and legacy customer service and knowledge tools. Besides process guidance with next best steps that reflect regulatory compliance and organizational best practices, modern KM systems also excel in findability, knowledge federation, personalization, omnichannel delivery, and proactive engagement. Moreover, modern agent desktops offer easy, push-button access to such guidance.

 

The Transformation

Here are some real-life examples in which knowledge and AI-enabled guidance delivered breakthrough results:

1. Telco giant achieved 30% improvement in NPS (Net Promoter Score), 23% improvement in FCR (First-Contact Resolution), and 100% improvement in agent speed-to-competency across 10,000 contact center agents and over 600 retail stores, while enabling any agent to handle any call
2. Multinational bank reduced agent training time by 50%, agent churn to <1%, and unnecessary service processes by 50%, while elevating NPS from #3 to #1, and enabling any agent to handle any call
3. Global bank reduced training time by 60%, improved FCR by 36%, and NPS by 10 points
4. Leading telco reduced unwarranted handset exchanges and returns by 38% through better resolution of customer issues by contact center agents and on the website
5. Leading manufacturing company deflected 33% of calls to guided customer self-service

All these clients had significant agent training programs but also thought that in-band search and process guidance for agents and customer self-service systems would be a great complement to training. They were more than right—the proof was in the results pudding, which was nothing short of transformational!