Off Script: Better, faster, cheaper – how AI transforms the economics of customer service
The long-standing belief in business has been that you can’t have it all. “Better, faster, cheaper – pick two” has long been the mantra when it comes to balancing quality, speed, and cost.
The same went for customer service. You could have swift, world-class support, but it would cost you a small fortune, or you could get a quick, budget-friendly service, but the quality would leave much to be desired. The trade-off seemed inevitable, forcing businesses to prioritize certain aspects over others depending on their strategy and resources.
Until now, because the advent of modern AI is turning this phrase on its head. You don’t have to choose anymore – you can pick all three.
In this episode of Off Script, our series of candid conversations with Intercom leaders about the extraordinary technological shift being driven by AI, Intercom President Archana Agrawal explores how AI is changing the economics of customer service and the implications for businesses.
Here are some key takeaways from the episode:
- The internet has raised customer expectations for great service, but businesses are struggling to meet these demands and maintain cost efficiency at a global scale.
- Cutting down on support investments may help reduce costs in the short term, but it won’t boost returns in terms of customer satisfaction, repeat purchases, or retention.
- Customer questions can be categorized by degree of complexity, urgency, and personalization – and this framework allows businesses to determine which questions AI can handle and which ones require human intervention.
- Businesses can adopt AI in stages, from basic task automation to full AI integration, allowing them to start small and scale up as needed.
- With the adoption of AI, CS teams will shift from reactive to proactive roles, becoming more consultative and creative, and taking on more revenue-generating responsibilities.
We publish new Off Script episodes on the second Thursday of every month – you can find them right here or on YouTube.
What follows is a lightly edited transcript of the episode.
Off Script: Episode 5
Archana Agrawal on how AI transforms the economics of CS
Eoghan McCabe: Better, faster, cheaper. Conventional wisdom says you can’t have all three, but we’re finding that modern AI is challenging that and that the new AI customer service systems will actually be all of the above and change the economics of customer service and digital business in fundamental ways.
In this episode of Off Script, Intercom president Archana Agrawal will take us on a journey through this change. She’ll share how AI has not only improved service quality, but may finally cause customer service to be viewed by some brave brands as an opportunity to drive value rather than simply be a burdensome cost. As we figure this all out together, I think this will be highly valuable viewing for everyone involved in the business side of online business. I hope you enjoy it.
From local service to global reach
Archana Agrawal: Businesses are operating at internet scale, catering to a global audience, but yet the pressure on them is to reduce costs, to have better profit margins. And that has become incredibly hard as these businesses scale. They always come down to, “Hey, can we shave a few percent points from customer support?” And that’s what ends up happening in trying to pursue this efficiency. Really, the question is: How much do you care about your customers and how are you going to invest in their experience?
Customer service has always been part of businesses, but there was almost a golden age of customer service where it was built into transactions, it was built into how businesses worked with their customers. And that’s because retailers and service providers were all local. People bought within their neighborhoods, within their cities, and such transactions were very, very personal. It was in the post-war era, I would say, when customers were spanning countries and continents, that businesses realized they actually needed a professional customer service function.
The idea that you needed a customer service team only happened when businesses started expanding beyond their local region, and it all started with the phone. Think about the call centers, the IVR systems, the 1-800 numbers – that’s when it started.
“Now that consumers expect speedy, accurate, and convenient service, it has become mandatory for businesses to deliver on that”
In the 1980s and 1990s, businesses started outsourcing their customer service to third-party vendors, and that’s when you had these globalized large call centers coming up in different locations. Then, maybe in the late 1990s and early 2000s, when CRM and customer databases became things businesses were adopting, that’s when email became such a dominant channel. After that, business messenger and social media channels have vastly improved the access customers have to businesses to get support and service when they need it.
With the internet, the way businesses provided service changed, and the way customers expected service also changed. The same technical advances that helped businesses move so fast made information available to customers easily. Customers were empowered, so they started expecting businesses to provide that information and support just as easily.
There were social channels that they took to in order to express their opinions and positive and negative experiences with businesses. As a result, that actually increased the pressure to provide better customer service. Businesses are operating at internet scale, catering to a global audience, and yet the pressure on them is to reduce costs, to have better profit margins. And that has become incredibly hard as these businesses scale.
The costs of cutting corners
Now that consumers expect speedy, accurate, and convenient service, it has become mandatory for businesses to deliver on that. It is how they can fuel their brand. It is how they can ensure retention.
This has been a paradox for businesses, right? Study after study, we know that investing in great customer service and treating your customers right leads to great business outcomes. On the other hand, businesses are also under intense pressure to be profitable. As they look at their profit margins, they try to figure out how to keep these as high as possible and how to reduce costs, and they always come down to, “Hey, can we shave a few percentage points from customer support?” And that’s what ends up happening in trying to pursue this efficiency.
“A poor customer experience cannot help; it always results in upset customers”
It is true that by pulling back your customer support investments by a percentage or two doesn’t mean that customers are going to leave immediately, but eventually, they get frustrated and in very public ways, they say, “Hey, what happened? You used to care.”
It has become very important for businesses to understand that it is not a trade-off of a percentage here and there. Really, the question is: How much do you care about your customers and how are you going to invest in their experience?
Around the 1990s and 2000, that’s when businesses started deflecting customer queries to reduce the number of direct interactions and control costs, but it actually held a number of businesses back when the methods they used introduced friction, such as making it hard to find how to get in touch with customer service.
In the early days of this deflection, it was the early chatbots, and then you saw IVRs and autoresponders. Sometimes they worked, and sometimes they created negative customer experiences when people couldn’t get through to the business to get their questions answered. When badly done, these deflection tactics, which may help control business costs, don’t help in the long run. A poor customer experience cannot help; it always results in upset customers.
Great customer experiences lead to great businesses
Many businesses have innovated in order to create great customer service. Think about Zappos – they put customer delight front and center.
Amazon has invested so much in great customer experiences and used technical innovations to do that. In fact, businesses today think about very customer-friendly refund policies. When they know they have repeat customers, they’re able to provide much faster refunds to those customers without hurting the business.
And then there’s Nordstrom, where they invested in making sure that whether online or in stores, customers feel like they get a seamless personalized experience. So, it’s possible to invest in your customer service and see the returns these businesses have in terms of repeat purchases, customer loyalty, and great customer satisfaction.
Tailored solutions for different queries
AI changes the economics of customer service. Now, it’s possible to provide great customer service that’s faster and even more cost-effective. Every question that comes into a business has some degree of value for the customer and the business.
But to understand how AI can actually change the economics of all of this, consider every question in three dimensions:
- Complexity
- Urgency
- Personalization
When you think about it in this format, you can actually see which questions we should have AI resolve, and where we need human intervention.
For low complexity, low urgency, and low personalization, we should have AI agents resolve these immediately with fast, instant answers. The human agents can spend their time working on more complex problems and customers get very quick responses that are accurate.
Maybe we have low complexity and high personalization questions, like, “I want to know my last transaction details.” If AI has access to your datastores, it can immediately read that transaction information and provide it to the customer.
Then, there are high-complexity queries. That’s where humans shine – empathy, expertise, where you’re able to use your trained human agents to answer customer questions.
The category of high complexity and high personalization questions, think about getting a customer to start using your product. Proactive support can provide personalized onboarding instructions to help a user get started, and frankly, that’s such a revenue-generating activity that I see support teams participating in the future.
Once you’ve implemented AI and all of these systems, that’s when you start seeing the metrics go up and to the right. You see improved customer satisfaction because your customers get faster answers. You have better profit margins, happier and more effective employees, and all the metrics we measure – COGS, FRT, CSAT – go in the right direction.
Cost effective for future scaling
While AI can be more cost-effective, it is probably quite expensive compared to what it will be in the future. That’s how technology has always evolved. Think about your personal computers and your mobile phones. Initially, they were hard to get, but then they became so affordable that they had vast adoption.
Similarly, businesses that can take the most advantage of AI today will be able to realize cost benefits today, but even as they scale, they’re not going to see their expenses scale like traditional methods. And so, finally, AI puts businesses in a place to provide high-quality, cost-effective, and very, very speedy service.
The AI adoption spectrum
Businesses can step into adopting AI. It’s not an all-or-nothing. They get to decide what day one is going to look like. Think about this as having multiple levels:
- Level zero is all human support and no AI.
- Level one is AI at a task level, focusing on low-complexity tasks that are repeated over and over, and we can automate some of these to help improve agent productivity.
- Level two would be workflow level automation and AI. You can take some parts or all of a workflow and automate it away or use AI in portions, and that can help improve the productivity of a team. If you use an AI agent today, like a chatbot, that can help answer some of the customer queries and deflect and route the more complex ones to agents, that would be workflow-level automation.
- Level three would be an exception level. Sometimes things don’t work and you need to have workarounds or exception handling, or maybe take a query and hand it to a different bot. That would be another level of adoption.
- Level four, the holy grail, where you actually use AI to help transform the way the entire customer service operates, like providing AI to help personalize the customer experience at every touch point, where you can completely revolutionize the way the function operates.
You get to choose and you get to start small and scale up.
The future of customer service
Customer service teams are going to be moving from being more reactive to more proactive. They’re going to be more creative and more consultative in nature.
You’ll find that there will be new roles created – roles that are working on knowledge content or AI conversation design. The adoption of AI is going to be a journey of continuous improvement. You will see a lot of training and upskilling in the process. It’s super exciting to think that, in the near future, customer service teams are going to be some of the first ones in organizations to be AI-enabled in terms of how they think and how they design their processes.
“Given that AI adoption is going to be gradual in organizations, it’s also going to foster a culture of continuous improvement”
As such, the function is also going to take on larger responsibilities. Now that they’ll be working on more complex tasks, they will also be stepping into more revenue-generating roles. We talked about proactive support, helping customers adopt products, and ensuring we provide guidance to customers before they hit any issues.
These are going to become very critical aspects of the role, and businesses will choose to structure the teams in the ways that make the most sense for their strategy. Some will lean towards white-glove, complete human support, that will be their differentiator. Others will move towards being able to automate a lot of their customer service through AI and automation.
Given that AI adoption is going to be gradual in organizations, it’s also going to foster a culture of continuous improvement. At the end of the day, AI agents must also improve. They’re not perfect, and human agents also have an error rate. But by focusing on how we can go through continuous improvement and work through providing the right content and context to the AI agents, we’ll improve the resolution rates over time.
Evolving customer service metrics
The way customer service teams measure their work will also change. For example, average handle time, a common metric today, is actually going to increase because the simpler queries are already being handled by AI, and the harder, more complex ones come to the agents.
A lot of the metrics may become inference-based, where you infer from the user behavior and their future actions whether the experience was successful. When a customer has a query and gets a response from the system, their future actions will help determine whether the customer had a good resolution or whether they needed additional help in any other dimensions.
By understanding more of the user behavior, we’ll be able to understand how good the service provided to the user was since they didn’t end up needing to make a service request at the end.
The trifecta of AI-first support
AI-first customer service is truly transforming the way businesses interact with their customers. Providing instant answers to customers that are accurate and cost-effective is going to have a transformative impact on the way they interact with customers.
It’s already happening. Businesses are already adopting AI in this way. You can start small, analyzing your queries, and finding where you would like to adopt AI. Start with answering queries on the weekend or for free users, iterate, improve, and fill the content gaps you have so that AI can improve. Hire specialists to help you design AI conversations.
“Businesses no longer have to make a choice between quality, speed, and cost-effectiveness. They can get all three”
Think about this: a small business today, by adopting AI, can actually provide 24/7 service multilingual service. Do you know what that does? That makes it go toe to toe with some of the largest businesses out there.
Today, AI can easily resolve a good portion of that business’s queries, but over time, it’s going to be 60, 70, 80%. It’s getting better and better at what it’s doing.
The future of customer service is bright thanks to the way AI is changing the economics and capabilities of support. Businesses no longer have to make a choice between quality, speed, and cost-effectiveness. They can get all three. AI-first customer service is now offering the trifecta of better, faster, cheaper support solutions.
