Getting Started with Generative AI In the Contact Centre

This article is featured on CX Today.

Generative AI promised to provide the spark that would light a fire under CCaaS innovation.

Now, with the first embers coming to light, few vendors are blazing a trail of ingenuity. Instead, almost every CCaaS provider has made the first same move.

That move is to develop a solution that auto-summarizes customer conversations.

How does it work? Well, after every customer conversation, it’ll create a transcript, redact sensitive information, and send it to ChatGPT – or another LLM – to generate a summary.

Agents can then copy and paste this into the CRM, cutting down on post-call processing, simplifying their workflows, and lowering handling times.

That drives a better agent experience, lowers queue times, and cuts business costs.

As such, the innovation gives end-users a quick and easy win. It also allows them to get more comfortable with generative AI before vendors release more ingenious inventions that may set the CCaaS world alight.

Yet, even at this early stage, some auto-summarization solutions are more advanced than others.

What to Look for In An Auto-Summarization Solution

One CCaaS vendor to release a LLM-powered auto-summarization tool is Cirrus Response. It’s the first capability Cirrus has added to its new AI Assistant.

Yet, despite it being the first capability, the offering has advantages over some of the other summarization tools flooding the market.

For instance, Cirrus allows the summaries to filter through to a central CRM, so agents don’t have to manually copy and paste them across to another platform.

“We call it client-side customization,” adds Ben Males, Revenue Operations Enablement Manager at Cirrus. “Users can go in, integrate their CRM with an API, and build their own post-contact flows.”

Another critical feature is the formatting of summaries. Each summary should have the same look and feel – states Males. That makes it easier for agents to handle future customer interactions as they flick through the previous call notes.

Moreover, supervisors, quality analysts, and coaches can more easily scour summaries, spot trends, and extract learnings to support agent development.

Finally, the CCaaS provider should allow the user to pull these summaries through as metadata to enable contact centers to extract further value from the solution.

Adding Value to Auto-Summarisations

Summarizing a customer conversation is not the only facet of an agent’s after contact work. For instance, they may have to tag tickets, send follow-ups, and enter data into different systems.

Working with a vendor that can mechanize most of these tasks and orchestrate a post-contact processing flow – which includes the automatic entry of summaries into the CRM – can shave precious seconds off every contact.

Now, consider the auto-summary metadata and how that could add value to various contact centre activities. Males gives the example of reporting:

“We already have tools to monitor customer intent. Combine that with auto-summaries, and businesses may highlight various trends across your most critical contact centre journeys.”

With conversational analytics on top of that, contact centres may also see how sentiment swings across contact reasons and conduct root cause analysis.

From there, businesses may fix issues upstream, apply contact automation techniques, and perhaps even rally for policy changes that customers clearly crave.

What’s Next On the Generative AI Roadmap?

The knowledge base is the next port of call for Cirrus’s generative AI innovation team.

Already its intent engine helps to inform the articles contact centers store there. Soon, the AI Assistant will scour those while monitoring customer conversations.

In doing so, the AI assistant will pull relevant content, combine that with customer data, and throw it toward an LLM. That will generate a personalized response for agents to review, lift, and send.

Many contact centres will already leverage agent-assist tools that work similarly. Yet, this added ability to wrap knowledge articles in customer context simplifies the agent’s role. Indeed, there is far less editing involved.

Yet, there is another element to this. As Males states:

“The AI Assistant may also flick through conversation transcripts across emerging intents and automatically develop knowledge articles for review.”

As a result, generative AI is enhancing the content that agent-assist tools leverage.

In addition, once those knowledge articles are approved, they may inform conversational AI projects, which widen the scope of the self-service mechanisms companies place on their website.

While this is on the horizon, it is only a taster of what will come.

“We’re also looking at other use cases and cementing more partnerships with LLM providers. One of those is Observe.ai – which builds contact-center-specific LLMs to support further innovation,” concludes Males.

Not to Fear, Cirrus Is Here

Not only is Cirrus considering the opportunities generative AI presents in expanding its portfolio, but the vendor is also leveraging the technology to extract further value from its current solutions suite.

As such, existing customers may harness its new AI Assistant to drive further ROI on current investments in conversational analytics, agent-assist, and automation technologies.

Of course, some will have reservations, such as: will this tech start to deplete my team, and is this technology secure?

Yet, Males assures us that the tool is about more than efficiency. Instead, it aims to lessen the load on agents and support their well-being – at a time when workloads are swelling.

As for security concerns, he states:

“We have the tools to redact Personal Identifiable Information (PII) and all other sensitive data, so any data customers train our AI Assistant on will remain anonymous.”

Cirrus will continue to take stringent security measures and keep a human-in-the-loop for all future innovations to avoid additional concerns – such as AI hallucinations.

With that seeping into the soil of Cirrus’s generative AI strategy, expect many more trusted, ingenuitive innovations to soon spring its portfolio.

Learn more about Cirrus and its new AI Assistant by visiting: www.cirrusresponse.com