In the OLD days
In 2019, before the pandemic, the only way to understand your customers’ sentiment was to send them an NPS survey, set up a conference call or get in a car or plane and meet them in person. The little information you gleaned was never enough, and If you were lucky, you came away with a health score of red, yellow or green that you were somewhat confident of, if not held hostage by.
Granted, two-years ago wasn’t the stone age, and, even by that time, a lot of work had already been done to try and make sense of the little sentiment information we did have. There were NLP models to look at the “free text” in surveys and deduce how a survey respondent felt about you and what they were concerned or happy with. But nothing really on the conference call front. In truth, even with a transcript, you were lucky to understand who was saying what.
And face-to-face meetings? Forget about it. I guess you could record them, but who brings a tape recorder to a face-to-face meeting? That’s so 80’s!
Fast forward two years, and the landscape has completely changed. There are so many more (mainly digital) ways to interact with your customers and just so many more conversations happening, especially with the increase to cross-functional teams. Add to that the volumes of operational outputs like usage data — logins, seat counts, features used — and you now have more information than you know what to do with or — in reality — make sense of.
The Sentiment Gap
In just two years, the post-pandemic world, and the digital transformation imperative it spawned, has placed us at the opposite end of the extreme. We now have so many new ways to connect with customers that it has generated its own multiplier effect.
More Channels X More People X More Conversations = Sentiment Gap
With all the conversations now occurring, it’s impossible for customers to not express themselves. The problem is no longer trying to get the customer to tell you how they feel about you, your company or product but capturing everything they’re saying — to you, your support, sales and executives, as well as to partners and each other — and making sense of it all.
With no way to filter and analyze all this chatter, the more customers tell us about their experience or subscription, the less you, the primary contact or CSM, actually know. This is what we call the sentiment gap and, unfortunately, without a way to close it, it grows wider with each conversation.
Up until now there has not been a way to coordinate these siloed conversations as they are collected separately from specialized systems. Instead, we surf through every support ticket, read through all of the comments and meeting notes in our timelines or rewatch our Gong or Zoom meetings at 1.5X speed — hoping that the complete picture of the customer and their sentiment will, like magic, materialize from out of it all.
Customer Intent Score from Conversations
With magic failing us, we turn to technology. But, even the latest advances in Natural Language Processing (NLP) are not enough, as the components are too specialized, and the most advanced models are just being made widely available. At CompleteCSM we created a Deep Understanding AI pipeline that houses the top technologies from AWS, Google AI, Cyrano and combines it with our own AI and ML components that we’ve derived from top scholars and their work at leading data science labs at Stanford, Harvard and MIT.
The AI models that represent the types of meetings are the target. Start with best practice meetings along the customer experience, modify them for your own specific customer Journey.
The key differentiator are the data processing engines and refine and process at each step of the way.
|NLP – Natural Language Processing Engine||Make unstructured conversations structured so that we can take it to the next step|
|Sentiment Engine||Look for positive and negative word choices|
|Emotional Engine||Find the emotional ranges by utterance of each person|
|NLP Objective Engine||Look at spacing between words and answers, questions asks, inflections, tone of voice and more for understanding|
|NLU – Natural Language Understanding||Understand what was said in terms of meeting notes, compassion, commitment and more|
|NLG – Natural Language Generation||Generate personality profiles, and how to respond and connect with people|
|Predictive Engine||Time series data and polynomial regression analysis to trend what is going to happen next!|
|Decision Engine||Insights on Churn and Expansion Analytics so you can take action|
It’s virtually impossible to have any intelligence on customers without the capability to analyze conversations. Not just support ticket status, but the back and forth conversation in emails of the resolution of the ticket. Not just a manual sentiment check on a customer meeting, but the deep analysis of how each person felt, and the meeting summary of what happened! Don’t forget about the Gong or Zoom online meetings that are so rich in customer sentiment that leaving them dormant would be a sin!
It’s time to step up your game and include full conversations in your customer intelligence strategy!
Learn more at CompleteCSM.AI and we’d love to chat with you!