Why Agent Assist Isn’t Successful

If you’ve landed here, you’re probably considering real-time agent assist something-or-other in your contact center. And it’s tricky to know where to start, especially with all the buzzwords floating around.

Just recently, I’ve heard terms like real-time guidance (that’s ours!), agent augmentation, call pops, agent assist, real-time agent assist, robotic process automation, and half a dozen of others.

To make sense of these cyborg-sounding phrases, I’m providing a human take on the agent assist conversation. I hope it helps you better research guidance technology.

How Humans and Agent Assist Work Together in the Contact Center

Agents are an essential piece of the contact center puzzle. Though some companies automate customer interactions, over 70% of consumers are either frustrated or angry when they can’t speak to a human.

This being the case, people aren’t about to disappear from contact centers. But even stellar agents need support to perform their best.

Imagine you’re given a goal: to reach the other side of an unfamiliar town within an hour. You could probably make it happen on foot, but a vehicle and a map would make it a lot easier.

The same concept applies to contact center agents. They know their goal and they can navigate calls without help, but useful information and support make the experience more seamless. That’s the role of agent assist — to supply agents with what they need so it takes less effort to deliver better customer experiences.

The Challenges with Agent Assist & Agent Augmentation

The primary challenge we see with many agent assist solutions is that they don’t meet the agents where they are. Though certain agent assist solutions provide ostensibly helpful information to contact center agents, they don’t really consider how agents should put the content into practice.

Again, imagine you need to make it across town within an hour. This time, imagine you’re given a car, a bicycle, a map, and a new pair of running shoes. But you’re not familiar with the area, so you don’t know which route to take or which mode of transportation to choose. You know your goal, you have the tools to achieve it, but you still don’t know how best to go about it.

Too often, agents are left in a similar position. Many agent assist products are almost entirely a voice-activated database search. Speech-to-text technology turns voice to keywords, then searches against a database for information with matching keywords.

With these products, somebody says “pricing” and a pricing PDF pops up. Agent assist solutions supply information, but they don’t tell agents how to use that information. Basically, agents are left with all the tools for success but no instructions on how to wield them.

Agent assist is uniquely give-and-take. It interacts with contact center agents frequently throughout their workday, but success hinges on agents wanting to mention the content presented. They must commit time and effort to find how best to incorporate that information on their own.

Without agent buy-in, information is useless. Simply surfacing content isn’t enough.

Think back to your best coach, teacher, or mentor growing up. Was their feedback thoughtful and relevant to you? Did you want more feedback from them or less?

Conversely, we’ve all had coaches and teachers who bombarded us with a ton of bad information. That’s the absolute worst.

Agent assist is the same way. Your success with agent assist depends more on the quality of the suggestions than the frequency of those suggestions. Infrequent yet helpful suggestions are far more useful than a steady stream of unhelpful suggestions.

If an agent assist solution surfaces information six times on a call, but it’s only helpful three of those times, will the agent still dedicate time to fitting the content into the conversation, or will they begin to tune it out?

On the other hand, how will the agent react if a solution only supplies suggestions three times, but it’s relevant and helpful each of those times?

Simply put, many agent assist solutions supply (lots of) information but leave agents to figure out if and how to implement it.

So, How Do I Make Sure Agent Assist Suggestions Help?

This is where the differences between “guidance” and “assist” come into play.

Guidance implies leadership and proactivity — guiding an agent through a challenge takes active effort. Assist doesn’t have the same implication. Assist is subordinate; it’s an auxiliary function that depends on someone else taking action.

In 2017, we founded Balto. Today, we’ve guided over 130M calls in real-time, so we’ve seen nearly every mistake somebody can make with agent assist and similar contact center software.

In fact, myself and my co-founder were sales development representatives (cold calls, outbound, Redbull…the whole shebang) before starting Balto. We were, and very much remain, not computer scientists or software engineers. We learned what works and what doesn’t through trial and error. Eventually, we landed on Real-Time Guidance.

Guidance considers the goal, the tools at hand, and shows how best to achieve it. Like a local guide who tells you the fastest way across town at this time of day is by taking a little-known bike path, guidance answers the question of how.

So how do you guide instead of assist? Proper guidance starts with experience. And who has the most hands-on experience? Agents.

In other words, great suggestions come from the ground up, not top-down. A solution should learn from agents and implement their best practices.

Unfortunately, most agent assist products are extremely top-down.

This flow of information is constraining, it’s impersonal, and it relies on an up-to-date database. Agents won’t adopt it because agents didn’t create it. Plus, the information takes up a huge portion of their screen, and it’s not necessarily relevant to the conversation. It’s not agent assistance, it’s agent annoyance.

Contact center agents often give the following agent assist critical feedback:

• It doesn’t drive a feeling of support or make agents feel leadership is investing in them.
• It doesn’t make the agent happier so that agents want to do their job more enthusiastically.
• It doesn’t acknowledge the skills of a tenured agent by recognizing or scaling their best practices to their peers.
• It doesn’t make agents feel autonomous, it makes them feel inhibited.
• It doesn’t make an agent feel rewarded, it makes them feel punished.
• It doesn’t drive resourcefulness, it drives monotony.

Effective solutions supercharge the individual agents rather than constraining them. The right products get the best out of the agents because they work as a force multiplier for the things agents do well.

Though contact center leadership has a better idea of big picture strategy, the best agents have a better feel on delivery. It needs to come in their voice.

No technology can force an agent to say something word for word, and why would you want it to? Their way of saying something best might not be your way of saying something best, but they’re the ones having the conversations.

Agent assist isn’t successful when it fails to take the final step of guidance, oversupplies information but under delivers on usefulness, and — perhaps most fundamentally — lacks the agent perspective. Make sure your agent enhancement solution is bottom-up, not top-down, or you won’t see the adoption or the results that you want.