Todd Pruzan, HBR: Welcome to HBR Video Quick Take. I’m Todd Pruzan. I am Senior Editor for Research and Special Projects. harvard business review.
In today’s customer service operations, artificial intelligence has become an integral part of responsible customer service leadership. Today we spoke with Brian Donahue, VP of Products at Intercom, to learn more about how and why AI agents can be your top support priority. Brian, thank you so much for being with us.
Brian Donahue, Intercom: Hi, Todd. Thank you for being here.
Todd Pruzan, HBR: OK, Brian, it seems like many customer service leaders are struggling with how to incorporate AI into their support systems. So with so many options, how do you get started?
Brian Donahue, Intercom: One thing I often think about with AI agents is, because they’re on the front lines of customer support, what can they solve before a human agent intervenes? This is usually the most common and simplest type of question. In fact, the easiest way for people to think of it is to think of agents as primary support.
Todd Pruzan, HBR: OK, can you talk a little bit more about what Tier 1 support is and what AI means for your operations at this stage? I mean, how will your customers experience AI? ?
Brian Donahue, Intercom: Yes, there is no standard definition of what Tier 1 support means. Many companies mean something a little different, or sometimes more specific. So our idea is to really categorize what kinds of questions we get. Therefore, one of these question types is likely to be an information question from the customer.
These are generic. The answer is general. It doesn’t matter who your customer is, the answer you’ll get is the same. This is the easiest place to start your AI agent. All the answers are somewhere in your knowledge base, or you want the AI agent to be able to answer them, or you should, so you just have that supporting content. It’s simple. Everyone understands that part. However, it is useful to think about it. This informational content may include, if your service or business is large enough, you may have product specialists, service specialists, and people who specialize in different areas. Because it’s more complex, deeper, deeper. It’s even harder for customer service representatives to truly understand the scope of your business.
However, in this case, the AI agent can scale across its breadth and depth. So, in reality, the AI agent not only handles the information questions, but also the specialization of its products. As long as the documentation is there, the AI agent will recognize it. Yes, this is bucket 1 information.
Todd Pruzan, HBR: information. That makes sense. So where do we go from there?
Brian Donahue, Intercom: So if the information is general, the answer is basically the same for every customer. Then the second bucket is when that’s not the case. It needs to be personalized. So let me give you an example of what typically happens here, where someone says, “Hey, when will my new credit card arrive?” Obviously this is specific to that customer. Answering personalized questions means that the AI agent needs to connect to third-party data (third-party from support tools). So it extracts the data, matches it to the customer, and says, “Your order goes here, your credit card goes here.” This is the basis of personalized data, or, sorry, personalized answers: extracting data and customizing that answer.
What’s really amazing is that the capabilities of the AI agent are amazing. You can give all this data. When a customer asks a question, this data is synthesized as needed and I can answer the question accordingly. So the way the answer can be synthesized is actually quite surprising. This is bucket 2 personalization.
The third bucket is actions. Personalization still only provides information. It is specific to that customer. The third part is the action part, which means doing something. It’s like, “Okay, okay, I’ve been waiting that long for a credit card, and the AI agent says, “Hey, I know you’re eligible for a virtual credit card.” is. Is it okay if I set it up? You can then move forward and the end user can essentially ask the agent to act on their behalf.
So this is where the agent can not only access information and synthesize answers, but actually initiate changes on behalf of the customer. That means hiring agents, canceling orders, placing orders, and setting up virtual credit cards. So that could be the third part of it.
What’s really interesting is that the ratio of information to personalization to action varies from company to company. Therefore, it can vary considerably from company to company based on the type of business they operate. But we basically think that once these three buckets are run by an AI agent, that’s really basically layer one support.
Todd Pruzan, HBR: I see, that’s amazing. So, does having Tier 1 support handle many of the basics mean your human employees can focus on bigger customer service challenges?
Brian Donahue, Intercom: Yes, that’s exactly right. This is actually kind of fulfilling the promise that technology was meant to make for customer service for years, and this is what AI agents are actually doing. Humans can take on tasks that are more difficult, tasks that are more complex, tasks that require real problem-solving skills, and tasks that require judgment about what to do here.
And when a customer comes in and sometimes they need some really human empathy, that’s actually the most necessary thing to turn them from a frustrated customer to actually a loyal customer. So this is the only place where humans have to be involved. This is basically the second layer. So it frees up the human from answering things that are simpler and maybe a little bit more sophisticated in the case of personalized things and actions, but it allows the human to really focus.[ed] Address the bigger, harder, more complex types of questions that customers actually need a human to solve. So it’s basically layer 1 support for AI agents. Human, Tier 2 and above.
Todd Pruzan, HBR: OK, that all makes sense. Brian, thank you so much for sharing your insights on using AI agents in Tier 1 support.
Brian Donahue, Intercom: Thank you, Todd. Thank you for having me.
Todd Pruzan, HBR: I spoke with Brian Donahue, Vice President of Products at Intercom. Click the link below to learn more about how to grow your support operations without expanding your team.
Click here for details AI agents and the role they play in responsibly serving customers.