AI supports contact center agents with Live Real Time Guidance & scripts, allowing live agents to focus on offering better customer experience.

By Colin Taylor

[This post was adapted from a webinar that Colin Taylor delivered for Customer Contact Central on September 8, 2022. You can see the replay here while available:]

There has been a continuous evolution in scripting since we were taping printed copies onto cubicle walls. There are primarily 4 different types of scripts that are employed in contact centers and many variations on these.

Traditional scripts

These are the type I was writing 35 years ago. These outline an “Ideal” path with suggested verbiage for the call. They were disconnected from any systems. Of course, the biggest drawback for any script was that the customer on the other end of the phone didn’t have the script and consequently didn’t know how they were supposed to respond.

Often, these reflected a checklist of all of the elements that needed to be addressed. These could be compliance related or behavioral. These checklists in turn became quality evaluation forms. These checklists and forms tend to be more structured for Outbound calling (Greeting>Qualification>Presentation>Objection Handling>Close>Objection Handling>Close).

Individual performance, where we expected it to be similar between agents using the same script, could still vary significantly. Often the agent’s ability to “make the script their own” was the key to success, as was the agent’s ability to recall and use their product knowledge and training.

Real-Time Guidance scripts or RTG

RTG introduced a process based on interviews, call recording analysis, and internal best practices to scripts. It included sub-processes and screenshots of the appropriate systems with links. This type of script, however, doesn’t have connections to any systems such as CRM, Knowledgebase, etc.

A detailed step-by-step process map is illustrative of the level of detail required in developing Real-Time Guidance. We need to show every step in every call; every decision, every outcome for every decision.. We need to include agent guidance to do X or Y, and we need to follow sub-processes where appropriate and not lose track of where we are in the process. While this guidance isn’t connected to any systems, it often includes screenshots, highlighting what needs to be added, checked, entered, etc. Knowledge articles can be embedded in the structure. This represented a great improvement over manual scripts.

It is not surprising that there can be a compelling Return on Investment argument for this form of Real-Time Guidance.  Studies[1] have shown that Real-Time Guidance can:

  • Reduce Training Time by 75%
  • Reduce AHT by 12%
  • Increase FCR by 6%

The most impressive result, in my opinion, is the 75% reduction in training time. By having the step-by-step guide on screen throughout the call the agent has constant support. It also means that they don’t need to know everything; they just need to know where to find it.

Real-Time with Dynamic scripting

This added to the mix connections to company systems, CRM, telephony, Knowledge, and Training platforms. When we take everything great about Real-Time Guidance and add in live system links, we have removed another barrier and eliminated an agent task. This simplifies the process and reduces the likelihood of errors or mistakes.

Live Real Time Guidance

Lastly, in “Live Real Time Guidance” we add AI and many features and enhanced benefits such as customer sentiment and next best action through “in-the-moment” guidance. This represents a quantum shift, as for the first time, customer sentiment can be assessed in real-time and guidance isn’t happening after the fact during a coaching session focusing on “what to do in the future.” It is happening now, and we can act now, in real-time.

AI can “listen in” to the call and act/react based on what it hears. We discussed checklists earlier, and these script guides have been around for years. The difference is that now the technology can track whether we did or didn’t address all of the items on the checklist. This can now augment or perhaps automate our quality programs. The agents can now receive “In-The-Moment” guidance with suggestions, prompts, etc. We can see the caller’s sentiment and offer corrective action, in real-time. We can send alerts to Supervisors who can then monitor the call.

Conversational AI

One specific form of AI that you may be familiar with is Conversational AI.

Conversational AI is the set of technologies that enables human-like interactions between humans and machines. Using machine learning, natural language understanding, and natural language processing, machines can understand how humans naturally speak or write.

This is the technology that enables smart IVRs, IVAs, and chatbots.

While there are hundreds of organizations leveraging conversational AI, there are two broad approaches: those that are automating the interaction via self-service, chatbots, etc. (called Digital Agent), and those that automate the process in support of a live agent. This latter variety is called Agent Assist.

Digital Agents, also called virtual agents aspire to automate and solve customer interaction issues and challenges without human intervention where possible. They listen, and recognize contact reasons, intents, sale opportunities, and the associated workflows. These can be deployed as chat or voice bots and in cases of low complexity, high volume workflows, they can be very successful in resolving inquiries. Up to 80-90% of the interactions can be resolved while the remainder can be handed off to a live agent.

Agent Assist deployments rely on the same or similar AI to digital agents. They can also listen and recognize contact reasons, intents, sale opportunities, and the associated workflows. They work to assist the agent by providing guidance, calling out omissions, accessing other systems, opening windows, and/or retrieving data that the agent will need. Live Real Time Guidance is an Agent Assist application

Live Real-Time Guidance

As discussed earlier, recognizing contact reasons, intents, opportunities, and the associated workflows plus sentiment, keywords, and phrase analysis has historically been the realm of post-call activity. This analysis could only be completed after the interaction had ended.

  • Traditional Interaction Analytics Historical- passive
    • Text and Speech analytics
      • Convert unstructured data for analysis
      • Provide future planning based on historical experience

Live Real Time Guidance is changing this pattern. It provides an open mic and can provide significant benefits while supporting the agent and allowing the agent to stay focused on the customer.

Live Real Time Guidance differs from the RTG we saw earlier from companies like ProcedureFlow, in which the step-by-step scripted process, annotated with screenshots and agent instructions, is disconnected from any actual call. Live Real-time guidance provides “in-the-moment” guidance based on the call as it takes place. The AI is assessing sentiment as it happens, based on tone, modulation, intensity, keywords, phrases, pauses, and silences. Research by Aberdeen showed that agents spend an average of 14% of their time just looking for information. Accurate live real-time guidance saves this time and eliminates pauses, holds, and hesitations, improving both flow and continuity.

The Live RTG can support compliance verbatim, disclosures, and caller authentication. Alerts can be triggered when a caller threatens the agent, is abusive, threatens legal action, etc. This allows a supervisor to monitor, coach, or join as appropriate.

Next Best Action guides the call and the agent forward. Live Real Time Guidance supports sales activities such as upsell/cross-sell and rebuttal. The agent can be prompted to show empathy, take ownership, be enthusiastic, build rapport, etc. In all areas Live Real-Time provides guidance on managing tone, pace, intensity, etc. supporting better interactions, more sales success, and improved CX.

And, because of its prescriptive nature Live RTG can integrate with existing coaching programs and practices by highlighting actual calls, and the Live RTG guidance. Similarly, because we know what the agent did or didn’t do, we can leverage the Live RTG data to support quality and gamification programs.

Once again there are some pretty impressive payoffs[2] with Live Real-Time Guidance. It can:

  • Improve sales by 26%
  • Improve CSAT by 25%
  • Increase Speed to Competency by 65%
  • Reduce AHT by 7%
  • Improve FCR by 6%
  • Improve Customer Retention rate

At the end of the day; or at least the end of this article; we need to ask,”So what does all of this mean to you and your center?”

Regardless of the approach you choose to take, there can be significant benefits to improving your scripting, and/or employing Real-Time Guidance or AI-enabled Live Real Time Guidance.

[1] ProcedureFlow

[2] ROI figures are from Balto & Cogito


Please Contact ApexCX for all your customer experience needs.

(Sep 28, 2022)