Balancing AI efficiency and human empathy in customer support: A CX leader’s approach
The discipline of customer experience (CX) has traditionally been very human-focused, prioritizing the personal touch. However, the current wave of AI advancements has seen automation enter almost every facet of business, including CX. This has led to a debate as to which CX elements can be passed over to AI before it starts to degrade, especially in the context of contact centers. That being said, many organizations that offer customer support want to reap the cost and efficiency benefits that the technology offers.
TechHQ spoke to Hosam Hassan, the Director of CX Transformation Strategy at leading customer support provider PartnerHero, to find out how AI is impacting CX in the post-ChatGPT world.
THQ: Can you provide an overview of the current landscape of AI tools in CX? How are they being utilized in the industry today?
HH: “There are two sets of tools out there right now; I think of them as ‘pre-ChatGPT’ and ‘post-ChatGPT.’ In pre-ChatGPT, they’re using an older technology where they were building, let’s say, custom models for each of the customer support teams that would be using the AI product.
“Today, they’re using more tools of the post-ChatGPT world where there’s less work for you to do to set up those tools, and there are several categories that customer support benefits the most.
“You have the agent-assist tools where AI is being used, for example, in a help desk to suggest articles to the agent, suggest responses, or translate things. These will still require an agent to be in the loop, but the agent is the one to finally send a response.
“Then you have the AI solutions that actually respond to customers directly, resolving issues without an agent in the loop. You don’t really have to write these intricate responses. You just give the AI a couple of bullet points of how to solve the issue and it goes ahead and writes its own message in a different language and gets out the response to the customer almost immediately.“
THQ: How would you advise an organization preparing for the integration of AI in its CX strategy, taking into account the potential challenges and opportunities?
HH: “I started my career as a customer support agent, and I could tell you right now the things that an agent would want automated are the things that they’re doing manually every day – the boring tasks!
“So start with talking to your team, find out what is super repetitive for them, what the boring tasks are, and see if you can automate those pieces first. Involve your customer support team and understand what they hate doing.
“For example, pre-AI, you would have a help desk manager and that person’s job is literally to receive all the tickets to read them, tag the ticket, route it to the right team, put a category on it, and then possibly send an auto-response. All of those things today, an AI bot can do. It can understand the intent of the customer, it can tag it with the sentiment, the category, and even route it to the right team. So that takes away maybe 50 percent of the work that an agent would have to do. All they have to do then is solve the problem and send the final response to the customer.”
THQ: How can CX professionals adapt their skills and approaches to effectively work alongside AI tools and technology?
HH: “Before AI, when help centers came out, agents would spend 50 percent of their time working on tickets and maybe 50 percent of the time working on creating self-serve solutions like help center articles and macros, which are canned responses.
“What they could do today is actually work on training the AI platform. The customer support agent is the most knowledgeable about the customer issues, so they should be looking at roles like chatbot builder, AI trainer, or prompt engineer.
“Leaders should get involved with the AI, make recommendations to the person building the bot, and see if they can take some of the work off their agents’ plates. Then agents can focus more on becoming more technical, maybe learn a new role or get certified in AI.”
THQ: Could you share insights on how PartnerHero is approaching AI in the context of CX and what services are being rolled out to support businesses in their managed CX transition?
HH: “In 2023, instead of going against AI and thinking of it as a competitor, we brought a couple of AI thought leaders onto our team, and we’ve essentially partnered with two companies that provide AI.
“What we’re doing now is offering not only outsourced teams to our customers but implementation of a Managed Help Desk, like Zendesk, Gorgias, and a couple of others.
“Then we’re also providing Managed Bots, where we can go in and build customers a bot from scratch. We understand their workflows because we’re already providing them outsourcing, and we’ll say, ‘instead of doing 100 percent human support, we found an opportunity to do about 20 to 30 percent through automation and AI, then the rest we’ll have escalated to our agents’.
“So we’re really embracing the change and evolution of the space and we want to be a resource for our customers, so they can trust us. We’ll take the weight off their shoulders so they don’t have to go learn how to do this or have to hire somebody new.”
THQ: How long would it take to build a Managed Bot?
HH: “It takes anywhere from a week to three weeks to get it up and running. With traditional products, you have to do some training on your existing data, but for newer products it’s generative AI. It’ll learn from your help center, it’ll learn from your help desk tickets, and you just have to guide it on how you want it to answer.”
THQ: Many businesses are interested in AI for CX but are concerned about the complexity of implementation. How does PartnerHero simplify the process of adopting AI, and what strategies are in place to guide clients through this journey?
HH: “We try to remove the technology from the conversation. I think that’s what complicates things. We start with understanding their use cases in the same way that we have to understand their use cases for outsourcing. We ask them, for example, what tickets are super repetitive, what the team does on a day-to-day that seems annoying or repetitive, and we would just map out those types of interactions. Then our team would say, ‘OK, we’ve noticed that we could quickly tag tickets for this customer; let’s start with that.’ Then, we build trust with our customers.
“We might not even send automated responses to their customers immediately, we would just start by tagging tickets. Then, once they trust that process and feel confident in the AI, we’ll implement something else. It’s more of a phased approach, we like to teach them along the way and show them exactly what we’re doing, so it’s not like a closed box or secret.”
THQ: In the context of CX, there’s a fine balance between human empathy and the efficiency that AI can bring. Could you share how PartnerHero addresses this balance and ensures that the introduction of AI enhances, rather than degrades, the customer experience?
HH: “Our approach right now, at least in this first year, is to build trust around the technology. We 100 percent believe in a hybrid model – we don’t think that 100 percent human is going to work any longer, but we also don’t think 100 percent AI is going to work at all.
“What we do is try to find opportunities to automate things that agents are doing on the back end that the customer doesn’t really know about. So tagging tickets, suggesting answers, translating. Then we leave the agent in the loop to do the final approval. That way they could add any final touches or they could recognize when a customer is having a poor interaction with the company and could stop it right there. We’re still in the building trust phase and we’re taking it very slow to get these tools up and running. ”
THQ: Can you provide examples of successful AI implementations in the CX domain where ParterHero has been involved and the positive impact it’s had on clients’ businesses?
HH: “With one of our partners, we’ve built an automation that essentially understands the intent of the customer and suggests the best article or the best answer for the agent. Now, the agent doesn’t have to go dig in multiple systems or jump around to different platforms; they get to just stay in the help desk. They get a little pop-up on the right-hand side saying, ‘Hey, we understood this ticket, this is the process that you should follow,’ and that speeds up the agent.
“Another example would be translation. If a customer contacts one of our reps in a different language that we don’t officially support, we can easily translate that through an AI plugin and then respond to that customer with the details they need to resolve it. It has definitely enhanced our efficiency and our skill set on our agent side .
“We aim for 20 to 30 percent increase in automation, so we’re automating 20 to 30 percent of the tasks that are being handled by an agent, and that’s on the lower end. We’ve seen examples of companies that do anywhere from 80 to 90 percent.”
PartnerHero is helping businesses design AI systems that solve their unique problems and take on the technical challenges that go with them. The company offers unparalleled in-house expertise from top software providers as well as access to best-in-breed chatbot platforms. Contact PartnerHero today to discuss how it can embed AI in your business and unlock the full potential of your CX.
22 February 2024
22 February 2024
21 February 2024