Get access to all the tools and information necessary for the customer's life cycle at Odigo by heading to our client portal
Natural language processing (NLP) is an applicable everyday contact centre technology. Discover how to harness it with these six practical NLP use cases.
As a field of study NLP brings together both linguistics and artificial intelligence (AI). This combination makes it perfectly suited to contact centre and customer service applications. Odigo, for example, has been using NLP in its contact centre solution for over 15 years. Deploying AI though shouldn’t mean running before you can walk. As discussed in the Get Out of Wrap podcast, using cutting-edge technology should first ensure businesses are getting the basics right. With this in mind, how can NLP help achieve this and take the next step to outstanding CX? Find out with these NLP use cases.
Many businesses believe they know the top reasons customers call. Often, however this is informed by post-call activities like wrap-up details. One very powerful NLP use case is identifying customer intent. For example, by introducing the question ‘What is the reason for your call today?’ at the start of an IVR menu real-life intents can be catalogued. In fact, within a few weeks, a great deal of accurate data becomes available. Initially, during data collection, any existing DTMF (dual-tone multi-frequency) options are best left unchanged. Afterwards, the more complete set of intents captured by NLP can be used to overhaul the functionality and efficiency of the qualification and routing pathways.
The likelihood is that most organisations will discover far too many possible call intents to fit into a DTMF IVR menu. Even sequential menus with multiple options would be slow, high-effort and frustrating for customers. This means that after the initial time investment, the next step in this NLP use case is to enable customers to simply state the nature of their query. This information can then be compared and matched with as complete a list of intents as a business can identify. Furthermore, it’s an iterative process which can be improved and enriched with new intents over time. As a result, qualifying queries can become quicker and more accurate, reducing customer effort.
Accurate routing, based on refined qualification is clearly easier, however, there’s more to it than that. With increasing query complexity and widespread upskilling and cross skilling, static routing rules simply can’t deliver high performance or efficient resource utilisation. Dynamic routing which can adapt to the conditions by utilising multiple distribution rules makes the best use of qualification data, agent skills and availability. Agents feel the benefit of career development and customer interactions are set up in the right way for first call resolution.
Personalisation is an important expectation for many customers but can only be achieved when you know who the customer is! Identification and verification (ID&V) is a crucial step in enabling more advanced self-service capabilities and convenience, but it can also prime agents for interactions. The availability of customer details prior to call connection along with the qualification details, informs agent actions right from the start of the interaction. A sequence of questions during qualification enables NLP to match customer responses to known details or answers to security questions. This means ID&V can be removed or reduced to a bare minimum during handling time.
Another aspect of this NLP use case is the potential for integration with specific AI applications that can add voice biometrics to the verification process. This small ID&V time investment pays off for customers and businesses by opening access to additional self-service features, VIP options or accelerating progress once they connect with an agent.
To some sentiment analysis may sound too cutting edge to be widely applicable however in customer service it’s not rocket science that phrases like ‘disappointed’, ‘contacting you again’ and ‘this is the second time’ signify unhappy customers. An important application for this NLP use case is identifying vulnerable customers, most commonly those who are financially vulnerable, with worries about bills or sudden changes in circumstances.
NLP qualification of these more emotive or delicate intents can also be informed by an organisation’s experience, typical queries, struggles or complaints that should trigger vulnerability checks or focused attention. A ranking or priority system can be considered to fast-track these sentiments where appropriate. Some savvy customers might not be above taking advantage of such a system, so contact centre managers would be advised to monitor this type of strategy.
Performance and satisfaction are central to contact centre success and there are clear NLP use cases for supporting agents in their daily tasks and ongoing professional development. It’s increasingly the case that agents deal with more complex and emotional subjects. Specialised training and skill sets are beneficial to customers, agent engagement and contact centre success. This can all be undermined though by poor working processes. If agents are routed calls they aren’t trained to deal with or if they don’t know why a customer is calling and have to scrabble to find the details they need, customer experience suffers.
Agents can tackle any query on the front foot when they are primed with customer context and relevant information. Integrations between CRM and Contact Centre as a Service (CCaaS) solutions, in addition to specially designed APIs, can as a consequence of NLP analysis of customer speech or text, offer relevant details as screen pop-ups. This could be to highlight resources in a knowledge base, relevant history or appropriate product and service offers. The result? More rapid resolution and added value for customers.
Self-service is an obvious NLP use case, bots and automation are enhanced with a greater ability to understand the intricacies of language. However, an upstream mentality can turn the rich data sets that come from NLP to the task of creating new experiences, not simply enhancing existing ones.
More accurate analysis of customer speech and text can enrich context with valuable nuances. For example, analysing the most common interaction intents may highlight a subcategory or request ideal for self-service but previously unidentified. This can point the way to novel touchpoints or experiences that deliver convenience to customers and cost savings for organisations through digital deflection.
For contact centres to get the greatest benefit from any NLP use case, well-planned deployment, rich data and project curation is important. Odigo can help create the optimal conditions for success thanks to a close partnership with clients throughout the design, build and launch of its solutions. Natively omnichannel, every interaction can be integrated into detailed customer histories and create more complete data sets. This, alongside multiple potential integrations and complementary 3rd party applications, means contact centres can equip themselves with a bespoke AI-driven set of functionalities for customer satisfaction.
If you’re interested in how to get the most out of an NLP investment, discover European executive experiences and opinions in the Davies Hickman – Odigo survey results.
Modern banking relies on in-branch standards of care and expert advice from a distance. Find out how Odigo’s Workspace app for MS Teams involves experts from outside the contact centre in the drive for customer satisfaction.
When details matter, Odigo’s Microsoft Teams connector can help maintain case continuity and accuracy.