Now that you’ve seen the potential for immediate ROI with artificial intelligence summarization, what other use cases come to mind? There are several that we’re very positive on at Thrio. But first let’s do a quick recap. We’ve seen how barriers to entry have fallen in terms of artificial intelligence adoption. The technology has become commodified and leading CX platform providers should include AI transcription in easily-deployable configurations. We certainly can do that at Thrio, and would be eager to show you our capabilities. But what’s beyond speech to text and summarization? One of the other key topics is broadly referred to as “agent assist.” This refers to a wide range of capabilities that are agent-facing rather than customer facing. And as we discussed a few posts ago, helping your agents be less stressed should be a key priority for you, as that can help reduce agent attrition. So how do we help agents?
One way is by reducing the number of other screens and platforms they need to go to in order to work. For example, a tech support agent may answer an interaction in a CX platform, but then toggle over to an internal knowledge base where they have to type an inquiry after talking to the customer to understand the basic issue. Then they’d have to toggle back to the CX platform to continue the interaction. This process might repeat itself several times as the agent and customer work to resolve the issue. Each time the agent switches screens, their stress level can increase, as can the potential for delays, mistakes, and increased handle time. We want to avoid all of this. So with AI-powered transcription, we can also feed that transcript to an AI-powered knowledge base repository. In practice, this means that your agents can skip the process of going to the internal knowledge base to look up potential solutions. The knowledge base, coupled with a modern CX platform, should be able to return relevant articles and resources to the support agent, right there in the original platform where the agent works. Again, this reduces handle time, increases first contact resolution, and reduces the stress on your agents. Positive, right? Well, there’s more.
With the right CX platform, the customer could describe their issue even before an agent gets on the call, so the knowledge base articles are already surfaced to the agent when the interaction starts. Combined with strong process automation in a CX platform, that agent could get customized scripting language that has them greet the customer by name, summarize the issue, and offer solutions right away. In practice that might look like an agent saying “Hi Jean, I’m John. I understand you’re having issues with activating your new mobile device that you received by mail last week. I’m here to help you resolve this issue as fast as possible. Do I understand the issue correctly? If so, let’s get started.” This has numerous positive effects– we’ve offered more personalized service to the customer that’s also faster and more efficient. The agent has had to do less screen-switching and can get right to what they do well, which is listening and solving issues. Everyone wins. And we’re doing it with the power of artificial intelligence that means you can deploy these capabilities with less difficulty.
Now, to be transparent: knowledge base integrations like we’re describing take more work than simply toggling on transcription and summarization. That’s why we don’t put these into the immediate-ROI bucket for agent assist. The reason is that your knowledge base needs to be formatted and ingested into the AI engine, and that can take a bit of work to execute. It’s much faster than ever, but it’s still not immediate. As always, we want to be clear and straightforward with you to cut through the noise and be realistic about what’s possible.
Other artificial intelligence use cases for agent assist are within reach once you have transcription. For example, tone and sentiment analysis can be returned along with the real time transcription. This can help agents better relate to customers in a given interaction. And for agents who may “inherit” a case later on, tone and sentiment analysis can arm them with insight about the customer and their unique needs. This can mean an agent has fewer surprises if they encounter a frustrated or angry customer, and the customer can hopefully reach an agent who’s prepared to be empathetic and understand their emotional state better.
Finally, another AI use case for agent assist is topic detection. An agent who “inherits” a case like we described may not have time to read an entire summary when starting an interaction, but AI engines can also return bullet-point summaries of prior interactions. And the right CX platform can add detail from other databases, like being able to offer a count of the customer’s recent interactions the customer so the agent can quickly understand that they had to call in multiple times. Again, this is all about offering better, more empathetic CX and doing it at scale so it’s efficient and repeatable.
So, what’s next? We’ve covered a lot in this series so far. Helping agents is a key priority, but we have other stakeholders in the business to help with AI as well. Next time we’ll look at how supervisors can be helped with artificial intelligence.
In our last post, we looked specifically at how AI-powered transcription can help your agents be more efficient by reducing wrapup time.