---

The AI-Assisted Experience Center

---

---

Drastically Cut Your Expenses and Improve Customer Experience with Advanced AI-Assist

Steve Jobs had a favorite anecdote to describe computers, and I find myself retelling it all the time. Studies have shown that the most energy-efficient creature on the planet today is… the Condor. And it’s not even close! Humans? We’re quite unimpressive as far as locomotion goes. Way down on the list. The Condor, on the other hand, can go an incredible distance with very little energy spent. It’s far and away number one… until you put a person on a bicycle. Then the Condor gets blown out of the water. In a nutshell, that’s Artificial Intelligence (AI)-Assist.

There’s a tendency to think of AI as a technology to replace humans. Anyone who has ever screamed at an automated phone menu knows that this is flawed thinking. Our research at Hunter Douglas has shown that you might be able to expect AI to be correct about 75% of the time. Not terrible, sure, but not reliable enough for a pleasant AI experience, either. Frankly, AI is still just not up to snuff. When you put AI in the hands of your agents, though, now you’re going somewhere! According to Forrester, companies that have combined AI with human agents have reported a 61% improvement in customer satisfaction and a 69% improvement in agent satisfaction.

Hunter Douglas is all-in with technologies that can make our agents quicker and more efficient—as long as we can justify the expense with longer-term savings. As we looked at AI-Assist solutions, it became clear that these savings add up quickly. Right now, it looks like AI-Assist will save us 12%, conservatively, of our required staffing in the future. And that 12%? That’s before you factor in added benefits with the way AI-Assist can integrate into our current data ecosystem. The real savings is going to be much, much more.

We’ve done the research and put in the work to determine costs and savings. And we’re thrilled with what this technology will do for our customer experience. Let me show you why we’re so excited about it.

Beyond Delivering Basic Information

Today, the go-to AI solution is to create automated systems that pull knowledge base information for your customers. The reality is you probably have this information available to the customer already, and if it’s not being looked at now, AI isn’t going to change it much. But if you get the right partner, you can take AI to another level. For example, AI-Assist can illuminate real-time friction points in your system. Say you’re seeing a friction point between the Customer Service Representative (CSR) and the customer. AI can flag the supervisor in real-time for the supervisor to immediately start listening and jump in if they need to prevent the call from going sideways. Now, you don’t have to find out about a poor customer experience when you receive a negative survey or get yelled at on Facebook—you can prevent it! For Hunter Douglas, these types of features are table stakes.

Real-Time, Automatic Quality Assurance

Another place we’re excited to see AI assist our agents is Quality Assurance (QA). You can have the AI do real-time coaching for friction points in the call, and then have it do real-time pushing for these QA situations. Picture this: you have an environment with a fair amount of legal requirements and somebody didn’t use a specific disclosure of verbatim you require at a particular portion of the call. The supervisor can be tagged and can interrupt before it becomes a breach or otherwise get you into hot water. How great is that?
Let’s take this further. At Hunter Douglas, every time we place an order, we have to do a readback of what we entered in the system so we don’t end up with a shade that’s built incorrectly and won’t fit in the customer’s window. Instead of using a person, auto QA can handle the task. Again, it won’t be perfect, but it’s going to be right 75-80% of the time and it’s listening to every call in real-time or near real-time (instead of the typical two or three calls per CSR per month). That’s way more data and way more intelligence on whether you’re getting the behaviors you want! And this data can point your QA team to a specific moment they need to pay attention to. AI might not be accurate enough to fix a specific situation, but it can tell you that it’s probably broken.

Use AI to Determine Sentiment

AI-Assist can tie into voice analytics. In fact, you should be able to apply it to any of your digital interactions (think all calls not just contact center calls, messages, texts, chats, emails, voicemails or video conversations company-wide). Sentiment finding is the key piece of that data. A true AI will help you find what’s broken so you know where, when, and how to address the important issues. How can you reduce your handle time in other places? What are interactions that are coming up that are an issue? These are quickly addressed with sentiment finding and AI-Assist.

What problems can you solve with this? Imagine a customer called in last time and there was friction, this time instead of an escalated customer with another random CSR, we’ll route the customer to someone with a better skill set to match the needs of the customer either technically, with emotional intelligence, or both and get ahead of the escalation before it happens.

We can plot CSR and customer sentiment in a matrix and identify patterns. For example, which CSRs are causing friction points? Or which specific customers causing them? Now you can clearly identify and build a strategy to solve it through routing, coaching, or some other form of intervention.

What about positive sentiment? We all should know which customers are more valuable based on their spend, size of business, and so on. Many organizations already deploy a strategy to give these customers a better experience when possible; here is another metric to get your most valuable customers with your best CSRs.

Auto Dispositioning Takes the Pressure Off the CSR

Today, most organizations rely on the CSR to tell you what happened on the call. And let’s be honest: consistency and accuracy are not where humans outshine computers. Auto Dispositioning allows you to not only take that lift off of the representative, but you can take out the CSR altogether. You can systemize what it’s looking for, get more accurate dispositions, and even get multiple dispositions at once.

One of our issues at Hunter Douglas is that the majority of our interactions are multi-issue. We’re a business-to-business company, so when a dealer calls us, they’re not calling us about one consumer and one shade. They’re calling us about 12 consumers and 15 shades. It is a heavy lift for a CSR. If we create a case for all of those needs and then uniquely disposition each one, we’re adding 8-10 minutes to the handle-time for a 12-issue call. That’s not feasible for any business! AI-Assist allows all of that to happen automatically.

AI-Assisted Managers

Often in a contact center, you run into calls that are spiking for some unknown reason. Let’s say it’s because your website is down. Typically you’d ask a handful of CSRs what’s going on. They’re going to tell you about their experience over the last couple of hours. Although they did get a call that the web page is down, it was one of many. So they’ll tell you it was a “mix of this and that”, but the calls didn’t spike because of all the normal calls—you’re looking deep instead of wide. AI-Assist allows you to see what’s spiking in real-time and, with auto dispositioning, it’s also telling you why. Now you not only have an AI-Assisted CSR, you have an AI-Assisted Manager as well.

All of the Data, None of the Data Entry

We’ve all been on a customer service call where we can hear the representative enter our information one excruciatingly slow keystroke at a time. It’s painful! AI-Assist can fill all of those fields, real-time. In fact, most of the data will already be populated when it recognizes your phone number!

How many times have you facilitated a meeting where you were the facilitator AND the participant versus just the participant? The conversation is way easier when you have a notetaker. That’s the next generation of AI-Assist. Instead of having to hire CSRs that have to have all these computer skills, you can hire CSRs that are simply really good with people. Today, half of the talk time of my agents is technical, administrative, and process. That’s just talk time! Add in call wrap-up time… it’s almost 90%. What could our CSRs do with all that extra time? And how focused of a conversation would they be able to have? Of course, the CSR is still going to have to pay attention to ensure accuracy, but the role is now completely different.

This one is a heavy lift to get going for sure. No vendor we’ve seen has a plug-and-play solution. It’s complicated. We have to create ties back and forth between our CRM and our AI tool, assign failure states and an ability for the CSR to validate that the AI heard the right thing, and handle our unique processes—even our own made-up words, like “duette”. If you’ve already done much of the above and are ready to take a swing at this—and this will become the gold standard in the next roughly 10 years—be ready for the in-house technical heft. Once properly deployed, your customers will thank you, and I imagine you’ll have a differentiator that will help you gain market share, reduce Average Handle Time (AHT), and improve accuracy—all of which will bring costs down.

Maybe, like me, you remember calling customer service in the 90’s and engaging in conversations where the CSR was just paying attention to you and working with you toward a solution. It’s becoming possible, using AI, to eliminate the administrative debt that prevents the CSR from fully paying attention to our customers—and bring the 90’s back… without the Hammer Pants™!

Conclusion

Done right, AI-Assist will make your life easier. You’re able to reduce staffing costs because you don’t need as many people. Your Voice of the Customer improves because the staff that you do have are more focused on the actual conversation. And you’ve systemized a lot of the things that couldn’t be in the past. Now if you need a new disposition? Instead of going out and training 200 CSRs and then calibrating them for months to get them to use it right, it’s just a couple of clicks for your administrators. Identify the triggers and dictionaries, then tweak it with the admin team. CSRs never have to be involved. Remember, it’s not about AI replacing your workforce, it’s about helping your workforce go further and do more with less, like putting a person on a bicycle.