💬 Stephen Jones

Sentiment Analysis is at the top of the list for tactics you can use to better understand trends in your customers' feelings about your brand. Getting actionable insights to empower your team starts with a leading tech provider like Sentisum, Chattermill, etc. - their AI will tag your customer conversations so you can TRULY understand what your metrics like NPS, CSAT, CES, mean. It's not good enough to know "Our Customer Effort Score is good, so we must be doing everything right". It's critical to understand *why* your customers are leaving the NPS scores they do, or rating conversations the way that they do 📈

One of the critical stages to implementing an automation tool for customer service is to ensure that the agent handover process is smooth from the perspective of the "end user" If your customer is talking to you on live chat, email, sms, social, wherever - they shouldn't have to do anything special to switch from your bot conversation to a human being It should live within the same chat window and the automated conversation should be passed to the agent to ensure that your customer doesn't have to repeat themselves You completely lose the value of the empathy that your humans can provide if bringing them into the conversation is a painful experience for your customer!

The most important thing is explore innovation in this area in a measured way Repetitive conversations and tasks can be automated (and they should be), but quality and quantity need to constantly be juggled! By taking a slower, measured approach and rolling out each use case in an iterative fashion you can make sure that you don't lose the human touch. AI should allow humans to be more human! Your CS agents shouldn't be sending messages that they're copying and pasting 🙄 But every use case has sub-categories which can be automated to greater detail in order to maintain a high level of customer satisfaction

One highly underrated value-add from automation is consistency. Templated responses across all of your communication channels make sure that the brand experience your customer's receive is consistent - however they decide to contact you! By identifying your highest volume contact reasons, and automating the replies you send for these, you can ensure that your brand's tone of voice, descriptions of products and services and policies & Standard Operating Procedures are strictly implemented. Relying on humans is always dangerous 🫣🤣

It can be tempting to throw automation at everything. E.g. I personally love using templates and text expanders, etc., and I never, ever manually write a full email anymore 😂 But there needs to be a plan! 1. Identify your highest volume use cases - using a tool like DigitalGenius's Automation Analysis 2. Map out your processes, clearly charting the flow of how your highest volume use cases are handled when they're handled manually 3. Slowly roll out use cases. Always make sure that the quality of your automated responses/actions is as high as your best manual actions 4. Stay on top of what is out there - there are so many options that you might never consider; like automating shipping carrier escalations, for example!

There's a lot of talk about 'Personalisation', but not not a lot of clarity about what's meant by that. There's a very close parallel to B2B sales; when you're cold calling and emailing a potential client there are two factors that matter most: 1. Timeliness - Why (are you contacting me) now? 2. Relevance - Why (are you contacting) ME? Does it matter which university this person went to? Probably not. It's the same thing with CX improvements. Collecting unnecessary data and 'personalising' their experience with stuff that is entirely irrelevant to what you're selling does nothing to enhance their experience. Stick to using their data to add context to "Why now?" and "Why me?" and it won't feel invasive!