Automated Transcript
I’m again joined by our resident expert, Martin Korner.
Hi, Richard.
Let’s jump in. Today we are going to get our geek on and go through tech. I know you love this topic because it’s like riding your bike. So, let’s talk about some of these technologies that are available that people can really use to supercharge their contact centres.
I think we should start with that topic on everybody’s mind at the forefront. I know you know ChatGPT. How are we going to leverage that in a contact centre, and how is it going to really start to supercharge customer experience for people?
Generative AI, especially, is going to absolutely revolutionise the contact centre industry. There are so many different opportunities and touchpoints where it can really help to improve the customer experience, agent experience, and make the whole process much more efficient and beneficial for the customer as well.
Could you give us some examples of that? Let’s dig into one specific area. Email, for example, is a very difficult channel because it’s long-form text. Customers can put multiple queries in the same email, and you can have HTML in there. It’s a difficult channel. What we can do with generative AI is summarise that email content. So, take long-form content and break it down to identify that the email contains two queries from the customer: query one about their order, and query two about another order from last week that they want to return. We can pull out other bits of information from the emails, such as order numbers and postcodes for identification and verification of that customer. Suddenly, we’ve taken this really unstructured piece of text, extracted key elements, and can use those programmatically. This provides opportunities for automation and makes the process more efficient by providing the agent with the information they need straight away.
So, you start to cut down on the time needed to read the email and figure out what the customer is asking for, using generative AI.
Exactly. As you’ve seen with ChatGPT, where you’ve experimented with it yourself, this isn’t complicated stuff to do. You can do it through a chat interface, structuring the right prompts, and then getting back the information you need to help the agent and the customer.
That sounds like a practical implementation of AI. Are there other subtle ways that AI can be used? When I think of AI, I imagine some big, grand system taking over, but what you described was AI working to help the agent achieve something. Are there other ways AI could do that as well?
Email sounds a bit old-fashioned for CX. However, with new channels that we deal with daily, does it work with them too?
Absolutely. I chose email because it’s the most complicated example and the hardest one to deal with. That’s an area where AI can give us the most benefit. But there are all these small touchpoints throughout the customer journey. For example, when a customer is on the phone or WhatsApp with us, we can analyse the sentiment of every message. AI can track the progress of that sentiment across the customer journey. If a customer comes with a complaint, it starts with a negative sentiment, which is expected. But as long as the sentiment trends upwards throughout the call or chat, and the customer is happy by the end, that’s a successful contact. We can apply this across all contacts, rather than just a small percentage for manual quality assurance. We can flag contacts with negative sentiment at the end for manual review.
Tracking sentiment in near real-time during the call allows for immediate action. If the call is going poorly, we can intercept it.
Exactly. A supervisor monitoring a team of agents can’t listen to all calls. But we can flag negative sentiment to the supervisor in real-time, who can then choose to listen in. We can also flag certain trigger words or actions in real-time, such as “I’m thinking of leaving” or “I want to speak to a manager,” and alert the supervisor based on the real-time transcript. The supervisor can then intervene in the call, rather than dealing with it later when the situation might have worsened.
We can analyse speech in real-time and take action based on what is being said. For example, if a customer mentions leaving, can we provide the agent with retention promotions and workflows to follow?
Absolutely. Surfacing the right information to the agent at the right time is crucial. When a contact first comes through, we should already know the reason for contact. We can collect that from the customer using voice. When the call continues to the agent, the knowledge base should already be open and searched by the query, providing the latest information. As the conversation progresses, we can surface different articles based on the topic. For example, if the topic changes to renewing a contract, we can provide relevant retention information.
That’s impressive. You’re describing a system where the agent is constantly supported by relevant information, improving the efficiency and effectiveness of the interaction.
Exactly. The benefit is that the information changes over time, and these knowledge base articles are always updated. Agents no longer have to scour the knowledge base; the right information is provided to them instantly. This allows agents to focus on customer service rather than searching for information.
This is great. Considering the high attrition rate in contact centres, as reported by CIPD, reducing the need for knowledge retention by agents can provide a more consistent customer experience. New agents can perform almost as well as experienced ones, reducing training time and improving overall service quality.
Absolutely. Reducing the difference in performance between new and experienced agents leads to a better customer experience. It also reduces training time significantly, allowing agents to start handling customer queries more quickly.
This is all very cool. We’ve talked about voice and email, but there are many communication channels like chatbots, web, messenger, WhatsApp, Facebook. Can similar techniques be applied across these channels?
The important thing is choosing a contact centre solution that is truly omnichannel, supporting all these different channels. This allows you to use the same technology across channels, building automation and AI into the journey and reusing it across different channels. Whether it’s voice transcribed to text or a text message, the same AI processes can be applied, providing a unified experience for the agent.
That sounds like technology is becoming ambient, enhancing experiences seamlessly. Are there other tools in CX that people should be tapping into, especially considering compliance and payment security?
Compliance, especially for taking card payments over the phone, is a significant challenge. Agents writing down card details can be a nightmare for PCI compliance. A product we offer, Smart PCI, completely removes agents from PCI scope. During payment, the card information is not recorded, and the agent cannot hear or see it. This supports fully automated payments and ensures compliance.
That’s great. You mentioned workforce optimisation and management tools. How is AI changing that world?
Any company not looking at AI for their products will struggle in the long run. The efficiency that generative AI brings is huge. When choosing products and partners, ensure AI is part of their strategy. It may not be in the project now, but it needs to be soon, or they will fall behind.
Amazing. Thanks again, Martin. Throughout this series, we’ve covered a broad range of topics. The recurring theme is customer centricity, providing the right information at the right time to deliver the best customer experience. Today’s discussion focused on choosing the right technology that can grow with the organisation, be flexible, and use AI to enhance both customer and agent experiences. There’s a lot to take from these sessions. Thank you very much for joining me.
