Enterprise Focus: Anticipating and Acting on Customer Needs
Derrick Bradley
This Thoughts by Andrei is penned by Derrick Bradley, NLX's Chief Growth Officer. Learn more about Derrick here.
It’s rarely the case for enterprises that customer data is easy to access, derive insight from, and act on in a short period of time. It’s pretty standard to have multiple teams and sets of stakeholders involved in answering questions like: “What drove the increase in volume to our call center last quarter? Why did our automaton rates drop last month? Where can we improve self service the most?” NLX is making it easy to answer these types of questions with a new set of powerful AI capabilities that we are busy incorporating into our platform.
We believe that enterprise customers who can quickly anticipate and act rapidly on emerging customer needs will develop a major competitive advantage in their respective industry. In an upcoming release, we will be offering enterprise organizations the ability to identify customer patterns and trends that may not be immediately apparent to them, in real-time.
This works by embedding our technology at the trunk of a smart IVR (voice) or a customer-initiated chat session so that we can capture and analyze a customer’s initial utterance and apply our zero-shot learning AI model to them to derive insights about your customers’ needs. The goal is to identify meaningful patterns and relationships with your customer data much like a librarian groups similar books together based on their characteristics, such as subject matter, author, or publication date to make them easy to locate.
- Audit your current implementation. We can quickly help you understand how well your existing implementation is performing, where the gaps are, and how to close them. This can be done without having to change any of your existing vendors or points of integration into other systems of record.
- Retrain existing intents. If a customer’s utterance should be connecting to an existing intent but isn’t, we can quickly retrain it with data that is grounded in what customers are actually calling/chatting about. For now these retraining recommendations are presented back to our customers for approval prior to upgrading the training data. Down the road we can easily imagine providing customers with the option to completely automate this continuous learning process without any (or limited) human intervention.
- Prioritize use cases based on customer ground truth. We provide real-time reports of what customers are inquiring about. This analytics capability is giving our customers the confidence to automate use cases that are driving the most volume or creating the biggest headaches for them and their customers. It answers both the “Where do we start automating?” and the “Where do we need to improve?” question. Customers are already finding this to accelerate internal decision making because it provides them with a fact-based approach to prioritizing the next use case to launch against their broader automation and customer experience goals.
- Build for emerging and temporal customer needs. If what feels like out of nowhere customers are driving more traffic to your call center and your smart IVR isn’t routing their requests appropriately, we can help you (i) understand why and (ii) resolve the problem to keep automation rates and CSAT scores in tact. For example, there may be an unrecognized intent for concert tickets that are suddenly available or an external event taking place (like poor weather) that is driving volume up. It may also be the case that a more general shift in customer behavior (like how they’re asking to cancel a reservation) is going unnoticed. Using our technology, you can be notified of these trends the instant they start to pick up steam.
The goal here is to help shift customers from being reactive to being proactive. This capability provides a level of insight that can sometimes take organizations months to reach. By providing this data in real-time, customers can meet emerging customer demands by building new voice, chat, or multimodal journeys (experience multimodal journeys here!) before they become call volume drivers and retrain existing intents with much better data. We’re already testing this capability with a handful of our customers and the feedback is overwhelmingly positive. It’s pushing us even further to reimagine the entire build-deploy-analyze cycle of bot building for enterprise towards something that is genuinely responsive to customer demand. More on that soon…