Conversational messaging is causing a disruption in the world of debt collection.
From low response rates to ignored emails, collections contact centres can get frustrated by processes that are both expensive and labour-intensive. But now, using conversational AI for debt collection is boosting customer engagement at a scale never experienced before.
Traditionally, companies have doubted the skill of AI chatbots to handle the tough conversations around debt collection and payments. More recently however, the technology has grown in leaps and bounds to the point that a chatbot can engage in intuitive conversational messaging that can deal with over 70% of customers’ issues. This allows agents to take care of the more complex tasks.
Multichannel conversational AI can revolutionise your customer engagement.
Conversational AI is technology that allows two-way communication with a computer program in a way that is both natural sounding and meaningful. Since it is conversational, people feel they’re being heard which is central to customer engagement.
Conversational tech, such as AI chatbots and virtual assistants, goes beyond the limited answer-tree rules-based chatbots and instead creates an automated back-and-forth personal conversation. The chatbot learns to understand the nuances of intent behind a human’s words and responds appropriately. Furthermore, an advanced bot can read the context of the words and pick out entities (such as dates and times) that inform its responses.
To move from simple responses to realistic conversations, programs must understand context and intent, which is only possible with better Natural Language Processing.
Natural Language Understanding (NLU)
A natural language is a human language (not code). NLU analyses unstructured text or speech and transforms it into a machine-readable format. Automated Speech Recognition (ASR) translates spoken language into a format that the program can work with.
Natural Language Generation (NLG)
NLG is how the bot replies to humans, i.e., generating the appropriate response. It uses Dialogue Management to create a response from its understanding of what is being said (intent recognition). The reply can be either text or verbal.
Natural Language Processing (NLP)
NLP is the two-way communication between human and bot (NLU +NLG=NLP). The bot correctly interprets written or spoken language and then responds in a meaningful way.
Machine Learning (ML)
In machine learning, algorithms use large data sets to learn and then predict outcomes. The learning builds up and accuracy improves over time so the program can perform tasks without specific instructions to do so. In language learning, the bot grows in its understanding of human intent which makes its responses more authentic.
Can you trust an AI chatbot to do the job?
An AI chatbot is good at dealing with straightforward tasks, however, as the machine learning grows, the bots become more accurate at interpreting human language. So, the more sophisticated the chatbots become, the more people trust them.
A blended approach works well where chatbots and AI virtual assistants work alongside agents to help them. If a request gets too complex, an agent will take over.
It is also important to have vigorous data protection controls in place to secure private data that customers share.
Propensity – Predicting the Outcome of a Conversation
It’s about context and intent, as Paul Sweeney, Head of Product at Webio, says:
“You need to create a context engine. Where am I in the conversation? For example, if someone says, ‘I can’t pay’ at the beginning of the conversation it means they don’t have the money to pay. If they say it near the end of the conversation, it probably has something to do with the payment gateway that’s not working. So, where I am in the conversation is context. And what’s gone before and what I am likely to do in that conversation – a projection of the outcome.”
Using Natural Language Processing, the conversational chatbot learns to predict how a conversation will go. Based on its interpretation, it guides you along the customer journey towards the desired outcome. It changes the way it responds depending on the unique situation of the person its talking to. (Learn more about how Propensity in AI works.)
Contact centres that rely on phones, emails and one-way SMS messaging find the following roadblocks stall the debt collection process.
Here’s how conversational middleware solves these problems.
Contact centres that switch to conversational customer engagement by using digital debt collection technology have the stats to back up their decision.
These are examples from customer feedback:
Reduced operational cost by 70%
Agent productivity up by 42%
Reduction in Agent Handling Times – 90%
Improved customer engagement by 52%
Positive outcomes up by 71%
Outbound messages outcomes – 58%
Conversations completed by bot – 40%
Customer engagement via chatbot – 70%
Response rate on proactive outreach – 70%
Response rate for Income & Expenditure – 60%
Other benefits include:
Do more with less
With automated debt collection, you need fewer resources to do the same job. Since the conversations are asynchronous, one agent can deal with multiple conversations at the same time. For example, an agent using a dialler can deal with 50-70 calls a day, whereas an agent assisted by AI automation can deal with 350-450 customer conversations a day.
Fewer phone calls; more conversations
Phone calls burn up resources at a rapid rate as they are so inefficient. On the other hand, using AI automation means fewer phone calls which means agents can handle more digital conversations at the same time.
Personalised and customer-centric
The data available on each customer - and provided by the customer - makes every conversation specific to that particular customer.
Faster and immediate
Since the conversation is automated, the process is quick and the collections chatbot provides immediate service in real-time with up-to-date data at hand. This also frees up your agents’ time to deal with the more complicated issues.
Agents feel empowered and agent retention improves
With the right tools and information at hand, agents can do their job better. API integrations help agents retrieve all the information they need to help the customer, with AI virtual assistants often dealing with this function.
Efficiency gains: automate the time-consuming manual tasks with a bot
For example, Identification and Verification (ID&V), Income & Expenditure I&E forms, account balance, etc. are simple tasks that the collections chatbot can sort quickly.
Better customer experience means better customer satisfaction
Your customers will be more willing to talk to you – and the outcomes will be better - if you make it easier for them. With conversational messaging, you can also include rich media, such as videos and documents.
Once the customer reaches the point of paying, unique links to payment gateways can be sent via the messaging channel, which makes it easy for customers to pay. Also, using conversational text messaging puts you ahead in the payment queue and customers tend to pay earlier in the process.
Simple to scale up
As your business grows you can easily scale up your digital messaging solution.
All debt collection laws and regulatory compliance can be coded into the application. This provides a secure audit trail that can be inspected if needs be.
Reports and analytics
Detailed metrics on how the debt collection software is performing is available, which also helps you to make informed business decisions.
Self-service is one of the most favoured benefits for customers. More and more, customers want the convenience of 24/7 instant access that allows them to connect at a time convenient for them and without the hassle.
Conversational self-service works well for uncomplicated tasks such as account and payment-related questions, balances, change of personal details, etc. An AI chatbot can stand alongside a customer and guide them throught the self-serve process and also answer any queries they may have. For more complex issues like payment difficulties and credit agreements, people prefer a speaking to a live agent.
Multichannel messaging (also called omnichannel messaging) is communicating with your customers across different channels. Conversational AI is used in multichannel messaging, such as two-way conversational SMS and other conversational text messaging channels like WhatsApp and Messenger. This is different from the WebChat widget you find on a website which is a live conversation with a human agent.
Since multichannel messaging is asynchronous, it allows customers to connect at a time and pace that suits them and gives agents the space to deal with more conversations at the same time.
Nowadays, customers want five to seven ways of interacting with a business. It’s worth it for businesses to do this as connecting with customers on their preferred messaging channel sees a 65% improvement in response rates.
Each generation has its own channel preference, with WhatsApp and SMS being the most popular with 25–34 year-olds while the 55+ age group still prefers phone calls. This is down to a general nervousness about tech in general among older people.
Customers don’t respond well to one-way communication as it sounds commanding. In fact, they avoid you if you’re demanding rather than relational.
Empathetic customer engagement
There is an art to designing a conversational AI chatbot for debt collection which emphasises understanding and empathy. To create this type of conversation, an AI bot needs to detect a person’s emotional state and intent and then guide the conversation towards a positive outcome. Thanks to developments in NLP, it is possible to have empathy and automation working together.
User-friendly and customer-first
By focusing on the customer experience, you will have better results. The challenge is to build intuitive collections chatbots that are easy to use for non-tech people and that talk to them in a respectful way.
Why customers prefer two-way messaging
Due to shame and anxiety, people don’t like talking about their debt. However, they feel better once they have had the chance to address it. A survey by Lowell shows that 69% of people don’t talk about their debt with over half saying embarrassment is the main reason. Conversational text messaging makes it easier for customers to talk about a way out of debt as:
Without a doubt, multichannel conversational AI for debt collection increases customer engagement, automation, and payments in a cost-effective way. The future of conversational AI is bright, in fact it's already shining as research by Gartner shows that the volume of interactions handled by conversational AI agents increased by up to 250% in many industries since the Covid-19 pandemic.
Chat to an expert in conversational AI for debt collection to transform your contact centre.
Extract from the Webinar "Conversational AI in Debt Collection How it Works Webinar"
For a more in-depth look at the How Conversational AI Transforms Debt Collection, download this guide that takes you through the 11 reasons why you need to go conversational in your customer engagement.