What Customers Really Think of Chatbots

Ready for a statement that will make you think? Gartner predicts that 85% of customer interactions will be managed without humans in 2021. This might be a scary notion for agents who are sensitive to job security but rest assured that human agent jobs aren’t going away. Why? The Contact Center is still inherently a forum for human engagement. The 85% of interactions that can be offloaded from human agents will be represented by the pesky repeat issues and questions that customers have that can sap an agent’s productivity. Imagine the additional productivity you can get from your agents if they aren’t chasing down account balances or processing payments? The reality is, bots and virtual agents will be a complement to their human counterparts, not a replacement. Bots and virtual agents will be helping their human teammates versus taking jobs from them. Chatbots and virtual agents have both become important spokes on the CX wheel – and that wheel will remain anchored by human agents. Used correctly, bots can become an agent’s best friend, helping them achieve and exceed personal agent KPI’s while helping lift the operational performance KPI’s, CSAT, and NPS that contact center leaders and agents live by.

NICE 2020 CX Benchmark

Each year, NICE conducts rigorous research on trends occurring in the market and shares those insights in the NICE CX Transformation Benchmark report. This annual report yields incredible insight into the minds of consumers and how their preferences are changing and evolving. This year’s report captures data points that reflect consumers’ interest in speed, efficiency, and the ability to solve problems on their own when the issue lends itself to self-service, like paying a bill, making an appointment, or asking a frequently asked question (see: things that bog agent productivity above).

What do customers think of bots?

The answer to this question is that consumers want bots BUT (there’s always a but) there is a twist. We asked consumers their opinions on using bots and other tech that enable them to solve issues on their own. While 84% of consumers say they are more willing to do business with a company that offers self-service options only 61% say that companies make self-service easier. In fact, when rating self-service channels only ONE THIRD are highly satisfied.

In other words, businesses have work to do in creating a satisfying, enriching self-service experience for customers. The first step is opening those self-service channels and give consumers the ability to solve issues themselves if they chose not to speak to an agent. When it comes to chatbots and/or virtual agents there are a few best practices to consider:

Know what problem you want the bot to solve. Don’t overtask a bot to handle too many things. Rather, create individual bots that each have an area of expertise. Bots are only as good as the information they contain, putting too much information in a single bot – or – asking a single bot to resolve one too many use cases will confuse it. Why? Because bots are taught to recognize keywords or key phrases. Multiple-use cases for a single bot may carry overlapping keywords, which will cause the bot to misinterpret the issue that the customer wants to resolve. That leads to an eye-roll and then a “can I just speak to a person please” mentality that leaves agents at a disadvantage. When the customer does connect with an agent, not only will they be frustrated but because they abandoned the session with the bot, the agent will also not have the advantage of viewing data collected from the bot to provide personalized service. This issue happening at scale can and will impact CSAT, NPS, and FCR among others, leaving YOU thinking “I made a bad investment.” You didn’t, you just made an error.

Bots are a workforce! Implementing chatbots or virtual agents is not a set it and forget it exercise. While virtual agents have the advantage of machine to machine learning, leveraging things like Natural Language Processing (NLP) and Natural Language Understanding (NLU) chatbots are a simpler being – both have a place on the CX wheel of spokes. For bots, ensuring that information is kept up to date (this may mean new products, promotions, processes, etc) is an absolute imperative. Depending on the velocity of change in your products and services, you may want a dedicated resource monitoring your bot and making updates.

ALWAYS ALWAYS ALWAYS, have an escalation path to a live agent at any phase of the engagement with a chatbot. I can speak (as I’m sure many of you can) to specific experience in dealing with this frustration. As a consumer, when presented with an “I can solve this myself” option – MY personal preference is to use that option if the reason for my engagement is simple enough to not require a human engagement. Once you commit to that though, it becomes incredibly frustrating if you are unable to resolve the issue yourself and you simply have to pick up the phone and call. Again, this doesn’t do your human agents any favors because the data that had been collected by the bot is not going to be presented to the agent. Making sure that the option to connect to a live agent in every stage of engagement with the bot goes a long way in keeping data intact, agents informed and customers generally pleased with the self-service options.

Bots, Automation, and Contact Center Results

It’s no secret that contact center leaders are looking for ways to accommodate increases in volumes, AI in its various forms in the contact center are tools to accommodate that, and guess what? They are good tools. Whether you are leveraging a bot or a virtual agent to help manage volumes and take routine issues off the shoulders of your human agents or leveraging AI-driven Workforce Management to help contact center leaders accurately forecast and easily schedule, the bottom line is that AI does two things, it reigns in the chaos and helps contact centers achieve their goals, it allows customers to help themselves and ultimately drives stronger connections, greater opportunities for repeat business and stronger customer loyalty. Leveraging technology like AI that helps customers help themselves improves CSAT, NPS, and FCR. Bots and other self-service options give agents time back in the day to focus on both helping customers with more complex issues and allows them to also focus on learning and development, upping their game and helping contact centers meet their targets, oftentimes exceeding them.

AI is taking a front seat in coaching

As I noted above with bots, know what problem you are trying to solve for and align the right AI-powered tech to support it. Looking to give contact center managers greater insight into operational performance? AI-powered analytics is a good place to start. Everything we’ve talked about with AI so far in terms of agent productivity and performance has revolved around giving agents back time in the day to either learn or to handle more complex issues that require human engagement. BUT – AI is also becoming an incredible tool for coaching agents in the moment, while they are on calls. Real-time Interaction Guidance delivers AI-powered insights to agents as calls are occurring – advising them on when they need to be actively listening, not interrupting, and taking ownership (among a long list of other behaviors.) This helps guide calls as they are happening, providing actionable coaching to agents so that they can course-correct in the moment, versus receiving feedback after the fact in a coaching session with a contact center leader. Real-time Interaction Guidance (at NICE, we call our solution Enlighten) represents quantum leaps in the application of AI in the contact center. Where algorithms essentially listen to the call as it’s ongoing and pops conversation bubbles (unobtrusively) to the agent to ensure that they are taking giving the optimal response. The result, at scale, is impressive – allowing contact center leaders to gain control over something that has, until now had to be done after a call with the customer has completed and a customer may have left the engagement frustrated. Having participated in some of the demos of Enlighten, this author can say first hand – that it is “Jetson’s-level” cool and worth checking out. 

AI Helps deliver knowledge

AI’s capabilities don’t end there, AI and machine-to-machine learning is also helping agents access information with incredible speed and accuracy as well. Imagine, if you will, that while an agent is engaged in a discussion with a customer and the customer asks about a product, or a specific process or policy, that an agent doesn’t need to go looking for it across disparate, disconnected information silos and systems. Rather, the piece of information the customer is asking about appears on their screen (just like the real-time coaching.) AI-driven knowledge bases are making enormous differences in delivering speedy resolutions to customers and are another example of how human agents and their ‘robot’ counterparts work together to deliver exceptional CX. NICE’s recent acquisition of MindTouch is bringing these impressive capabilities to the CXone platform. We believe this is an under-represented capability in CCaaS platforms and that it will deliver customers a meaningful capability that will help them differentiate their contact centers in this experience-driven economy.

Back to Bots

I’ll step off my “Benefits of AI” soapbox and get back to what you came here to read about, chatbots and virtual agents. I’ve spoken with many contact center leaders over the last year and one thing that always comes up is the question “do you have recommendations on where to start?” Yes! I do. First, I will tell both Google and Microsoft have exceptional developer tools that allow you to build a chatbot using their respective frameworks. At NICE, CXone allows for BYOB (sorry not that one) Bring Your Own Bot, into the platform and use seamlessly. 

Next, know what’s on your bench. Data analysts understand the architecture and analytics – what’s more they can support integrating disparate data, structure or categorize data, support workflow and intent mapping. Assess your previous interactions and documented processes. Knowing those two elements will help you understand whether you can insource the development of your bot or whether to go to a Managed Services Provider or Systems Integrator to help you build your bots and virtual agents through an outsourced model. 

Visit NICE’s AI Educational Resource Page for a robust library of how-to’s and best practices to help you pick the right bot or virtual agent and explore AI and how it can help your contact center. You’ll find several eBooks which include a Starter’s Guide to Building Your Bot, best practices, and my favorite is our newly published , it’s a fun, easily understandable read that answers all the questions you wanted to know about AI in the contact center but were afraid to ask.

A few key takeaways. Bots and AI are not the enemy of the agent, they will help them emerge as stronger, more capable stewards of the brand experience and more reliable and trusted advocates for your customers. Next, if you are considering bots and virtual agents, it means you are also considering AI – don’t turn a blind eye to what other functions AI can help accelerate and improve in the contact center – things like workforce management (WFM) off contact center leaders a far superior means of forecasting and scheduling than the old tried and true method of pouring over spreadsheets and losing sleep at night. Additionally, consider what AI can do to help your agents in the moment with AI-powered interaction guidance and intelligent knowledge bases, both of which can put critical information on the agent’s screen without them having to hunt for info or wait for a post-mortem coaching with their supervisor. AI has emerged folks, it’s a mature technology and only advancing faster, make sure you take advantage of it.