In January, prominent CX futurist Blake Morgan predicted that 25 percent of AI agent-assist deployments will fail in 2024.
One critical reason is that many contact centers cannot unlock the necessary data or discipline to truly benefit from AI. That extends beyond agent-assist and across the whole spectrum of contact center AI.
Zeus Kerravala, Founder and Principal Analyst at ZK Research, previously made this point in conversation with UC Today.
He said: “Some companies want to connect their communication data with their CRM data… but how many companies do you know that love their CRM data?”
The unfortunate answer: not many.
However, in recent years, contact centers have started to utilize AI as an input mechanism to push past this problem.
“It could listen to a call, summarize it, and automatically update a CRM record,” continued Kerravala. “This could be useful in contact centers, sales, or customer success.
“With good AI, you can generate better data, which leads to better AI in the future.”
In stating this, Kerravala lifts the curtain to the intelligent contact center of the future, which leverages AI to bolster its data sets. It then utilizes those enhanced data sets to improve AI. That’s a powerful cycle!
Moreover, that intelligent contact center could go beyond the CRM example Kerravala gave.
After all, across the CCaaS space, there are already examples of that powerful cycle in action.
AI Training AI: 4 Fabulous Examples
No CCaaS provider can currently execute on all of the below. Nevertheless, each example showcases how an intelligent contact center platform could utilize AI to generate data and insights for other AI models to thrive on.
1. AI Performance Insights Inform Contact Center Routing
For years, CCaaS vendors have developed predictive routing models. These models analyze contact center data to predict which agent is most likely to deliver a particular outcome – such as a high CSAT score - for the specific customer reaching out.
Now, vendors can take this to the next level with AI-augmented QA systems – which surface new agent performance data across all customer conversations.
Indeed, their intelligent contact center platforms could scour that automated QA data to uncover which agents best handle specific queries.
That QA data could then inform a triage system, which routes contacts based on the likelihood that the agent will solve the customer’s query.
2. AI Knowledge Management Enables Next-Level Agent Assist
The latest AI agent-assist models leverage the content within a contact center’s knowledge base to draft customer replies or recommend next best actions.
However, there are often gaps where there is no knowledge article related to the customer’s query. Other times, a relevant article is within the system but outdated.
As a result, agent-assist models may supply agents with incorrect information. Some may even “hallucinate” and make up information for which the contact center can be found liable.
Thankfully, new natural language processing (NLP) and generative AI (GenAI) models can spotlight knowledge improvement opportunities and even draft new knowledge articles for review and publication.
With such tools, the contact center can reimagine its knowledge management strategy and ensure its virtual assistants leverage the latest and greatest knowledge base insights.
