Often times in the world of customer experience we see organizations implement technologies and processes to merely survive the current operational state. The problem with this strategy is that managers and directors have very little time left over to plan for innovation. Conversations in planning meetings often include phrases like “if only we had bots, then our operation would...”, or “if our customers could find the solutions to all their problems on our website, then we could focus on more important issues.”
The challenge is, every customer is unique. While they all share the same goal of entering your system wanting to solve their problem as quickly as possible, the similarities end there. I’ve been guilty of calling customer service and saying “representative” as quickly as possible, feeling like my issue was way beyond the IVR’s capabilities to solve with self-service. Was I right however? The reality is, if we are confident there’s an intuitive way to solve our issue without speaking to a representative, most consumers would will choose that route. I wouldn’t have cared if I was speaking with a chatbot or searching their knowledge base for the correct process. What I cared about most was solving my problem in an efficient manner. So, the goal then becomes designing multi-channel customer service solutions that instill confidence in the consumer up front that your solution, regardless of channel, is capable of resolving their problem without human intervention.
How can we implement solutions to reduce inbound load on our agents?
Here are some thoughts on ways we can minimize the load on our call centers by up to 50% in 2020.
Define Why Your Customers Are Calling – This seems so basic, but in today’s digital age this can be taken to a new level. With machine learning that harnesses natural language processing (NLP) or natural language understanding (NLU) we can read free form text and draw significant conclusions as if we were reading it ourselves. From an AWS perspective, this can be done with Amazon Comprehend. Using comprehend you can ingest unstructured data to discover insights and relationships in text. There is a treasure trove of potential sitting in your unstructured data. Customer emails, support tickets, product reviews, social media, even advertising copy represents insights into customer sentiment that can be put to work for your business.
The key is using machine learning to eliminate human error. Sure, we probably have an idea of why people call, but making sure it is true keeps us focused on solving the real underlying issues in our processes.
Once you have your reasons defined, categorize them into what you want your agents resolving, and what could be resolved by a self-service channel. Here are a few simple questions to get started:
- Is there a defined process to resolution?
- Is this information easily accessible?
- Will deflecting customers to self-service impact their experience negatively?
- Is this something you want your representatives using their time for?
I had a customer look me in the eyes one time and tell me that 60% of their issues were simple requests like reset passwords and locked systems. Do you really want your highly trained representatives focused on simple issues like password resets? In this specific case, they weren’t looking to reduce labor force, instead they wanted their agents focusing on more technical issues.
By ingesting your ITSM or CRM service history data into an application like Comprehend, you should be able to gain invaluable insights into what the real issues are, enabling you to build service channels that rely on proven data rather than educated requirements. Give this a try using Amazon Comprehend and see what comes out. You might be surprised. Then, once you get your new process in place, make sure you reevaluate your call reasons periodically to see how things change. Before you know it, you could be predicting your customer’s behavior and having frictionless solutions in place to resolve those common issues.
Promote Self-Service – Self-service seems like a double-edged sword. Yes, you want your customers solving their own problems, but their path to resolution still needs to feel frictionless. You can’t just put information on your website and assume every customer can find it. If this is the assumption, you will end up with frustrated customers wasting more time than they would if they had just called in the first place.
I am one of those customers who wants to solve his own problem and resolve them quickly. However, I need the confidence that the company has a quick path to resolution. Here are a few options that can add service channels to facilitate this:
- Chatbots – Chatbot companies have mastered the art of deflection. Once you’ve identified your most common issues, you can decide which ones are compatible with a clear path to resolution. Then you can create chatbots that consistently perform a predictably repeatable action. Amazon Connect can handle this with a Lexbot. Based on the same technology that powers your Amazon Echo and Alexa, Lex continuously improves over time given the amounts of data ingested every day. If you’re not an Amazon believer there are plenty of other options out there that can achiever similar amazing results with contact deflection.
- Improve Your Website Navigation – Hire a good marketing agency and make sure your website is user friendly and easy to navigate. Page leakage is oftentimes a driver for higher call volumes. This is an obvious and sometimes frustrating truth, but it helps to reduce the number of contact telephone numbers displayed and push more self-serve options. If your self-service channels are intuitive and useful, then your customers won’t mind the fact that you have buried your customer service number on your website.
- Knowledge Base – Make sure you have the answers to your top questions in an easy to follow and easy to discover format on your website. If your customer can’t easily find the answer to their questions, they’ll end up calling in more frustrated from wasting time. You can purchase COTS solutions to accomplish this, or build your own in AWS.
- Search Engines – Once you have a strong knowledge base, a strong search engine to unlock your content is key. This can help your customers find information stored in your knowledge base. AWS recently launched an enterprise search solution called Amazon Kendra. Kendra is a highly accurate and easy to use enterprise search service that’s powered by machine learning. Kendra delivers powerful natural language search capabilities to your websites and applications so your end users can more easily find the information they need within the vast amount of content spread across your company.
- Intelligent IVR – Direct customers to other channels on IVR hold messages. By pivoting an inbound call to a digital interaction, Intelligent IVR allows you to surface all of your digital assets, in a single location, to your voice callers. This dramatically increases the likelihood that your customer will adopt and use your various digital assets to effectively self-service their needs. These options should not be forced on your customer, as it is important to respect the customer’s channel of choice. However, signaling their availability may help lower call volumes in the future. One common solution now is to send the customer a text that engages them in chat so they can multitask. OmniLegion can build these solutions based on your needs within Amazon Connect. If you can dream it, there’s a good chance it can be built in AWS.
Omni-channel – Omni-channel continues to be a big buzzword within the industry. You should always promote different channel options while being careful not to force the customer down a specific channel. Deflecting to other options can be significantly less expensive assuming you solve their issues within that channel. You shouldn’t focus on any one specific channel and make sure you promote them all equally. Today, we see voice, chat, social media (Facebook, WeChat, twitter, WhatsApp, etc), email and SMS as the primary means to communication. Make sure you futureproof whatever solutions you go with and they have a solid roadmap including all or most of these channels.
First Contact Resolution – This may seem obvious, but you need to commit to first contact resolution. When you’re analalyzing your contact center data it can be easy to only focus on ways to reduce talk time. However, this may push your representatives into bad habits which can foster a culture of bad customer service. The miss here is the goal of resolving issues on the first contact, not reducing call time. You are working towards a frictionless customer experience, not a metric. This doesn’t mean you open it up to your customers having social calls. We’ve seen this with some organizations where people call that are lonely and they end up staying on the phone for 20+ minutes chatting about grandkids. Have the right balance of urgency and being human.
We are in a technological age where we can reduce or even eliminate human error in our analysis. You should never assume you understand all of your problems, or even that you can rank where they stand appropriately. Leverage solutions like Amazon Comprehend and other NLP / NLU services to take the guesswork out of it. Once you know where you really need to focus, dive in and drive the correct technology and process solutions to deliver the best outcome for your organization. Consider some of the newer AWS solutions like Amazon Kendra too. If you do this right, you could find yourself driving out half of your call volume enabling your company to grow without focusing on unnecessary HR needs.