How do you discover the best path to agent performance in the contact center? Right now, your agents might follow a training regimen developed through careful attention to past experiences and current performance. Or, they might be rewarded for wins and reprimanded for compliance failures. Whatever your approach, it (hopefully) includes information from the past to inform your future.
Even with a “path” in place, when it comes to following the flow of information, failures abound. For your average agent, 9 out of 10 customers hang up before a right-party contact (RPC) and only 1 to 3 out of every 100 customers actually end up paying — or even expressing an intent to pay. That’s a massive failure rate relative to your end goal (payment collection) and to the actions required to get there (starting with RPC),
Your top agents, however, are likely to reduce that failure rate by half. This performance gap leads to unsatisfactory payment rates as well as compromised customer experience and high compliance violation rates — none of which sound like the best path to improvement.
So, what’s missing? If you’re already generally using data on past conversations to improve current and future conversations, how could you narrow the performance gap further? The answer lies in providing your agents with suggested actions — themselves the product of clarifying the right path.
ProAssist, our new real-time agent assistant product, leverages deep-learning and AI algorithms to generate automatic cues to prompt agents on the best action they should take next. Your agents will navigate even the most complex situations by following the real-time prompts our product generates.
And how do we decide what those prompts should be? How do you know when an action is “the best?” We were inspired by both the purpose and the technology of Google Maps: In order to find the right direction, you need an end goal and some constraints. ProAssist uses that mechanism to hand-hold each agent in your organization so they can easily drive a call to success.
Here’s how it works: First, we use conversational intelligence to define points A and B, and then we discover the shortest route between them. We then map that journey from A to B into identifiable junctions, showing places an action could occur (like a left turn!).
Next, we constrain that route with industry and conversational guardrails most relevant to your specific work, and again find the shortest path. On step four, we implement compliance constraints and find the shortest path once more. Lastly, iteration — the model gets better and better and your agents are directed to through the best path to their destination just like they’re on a leisurely cruise — no deep thought or stressful decision-making necessary.
Don’t worry: ProAssist won’t drive your agents into a lake, à la Michael Scott. Let’s take a look at one of Prodigal’s clients as an example of success.
They hit basically every metric.
After using ProAssist, the average payment rates of those agents improve by 8.7%. More importantly, a far higher number of agents came extremely close to the top performers in collection rates after using ProAssist.
As far as a metric that supports those rates goes, that’s covered, too: Agents can cause 23% more customers to confirm an RPC and proceed to offering payment options.
Agents also see compliance shifts: They’re 20% less likely to forget about giving a QA disclosure or a mini-Miranda disclosure and therefore, they more doggedly adhere to industry and firm policies.
Active engagement with customers is another indicator of success. After using ProAssist, agents are 10% more likely to use an appropriately high level of empathy and become more confident with tackling complex customer situations — having been prompted with the best practices available.
If you’re ready to see how this innovative approach to finding the best path can help you reach your destination, reach out for a demo. Try ProAssist today to increase collection revenues, improve customer satisfaction, and minimize compliance risks.
.svg){kind=icon}
.jpg)
.png)
.png)
.webp)
