Tips for Scaling Customer Service Without Losing Touch

One of the greatest challenges for any growing company is scaling the relationships they have with their customers. We all love the intimate and personal nature of the local business customer experience, and that kind of one-to-one contact makes it easy for customers to provide the feedback that all founders crave, particularly in the early days when it can make a marked improvement in their product or service.

The mom-and-pop feel is easy to offer when you have five customers, or even 50. But as your customer base expands, it's exponentially difficult to maintain. With each new growth milestone, founders become further removed from that source of information and insight; the feedback flows through increasing layers of technology and people. Eventually, it bubbles up to someone whose job isn't to learn from it, but to reduce it, which might serve a short-term purpose but doesn't help the business in the long run.

At some point, all startups reach a tipping point where customer interactions cease to be an opportunity to improve and instead become a cost to manage.

From a business perspective, reaching that tipping point is a good thing. It means you have such a large pool of customers wanting to talk to you that you need to find a way to manage it. But growth does bring challenges. Live customer service is costly, and traditional channels like email and voice might not scale as quickly as you need them to.

As interaction volume grows, so too does pressure on agents. When they're inundated with a never-ending stream of tickets, collecting and reporting on customer insights can easily be deprioritized. The feedback loop that previously helped identify opportunities breaks and customer service quality declines.

The good news is that with a solid plan for growth and the right application of artificial intelligence, companies no longer need to choose between scaling customer service to meet demand and maintaining quality. AI can quickly and accurately understand and analyze heterogeneous data compiled from different sources (such as customer service transcripts or feedback surveys) at the speed and scale that's needed in today's fast-paced world, especially during periods of rapid growth.

With AI generating reports on customer insights, the feedback loop is no longer at risk of breaking and founders can easily reconnect with the data that helped them rise to success.

AI is already making a significant impact across many industries, and customer support is high on that list. The potential is massive, as our analysis shows that more than $500 billion in agent labor will shift to AI in the next decade, but success relies on execution that's thoughtful and purposeful.

The latest buzz around generative AI might make it seem like a silver bullet for CX, but keep in mind that it's only a tool. Putting it to good use requires a solid foundation of company knowledge and policy paired with strong application. Take care to set smart parameters for choosing technology partners in this emerging and exciting space.

Automated Resolutions: The North Star Metric

During the past decade, AI-powered automation has changed the way companies think about and deliver customer service, and success has typically been measured with containment (i.e., how many inquiries were not escalated to a live agent). While containment has its place in a CX leader's dashboard, it's a very narrow view of the efficacy of the automated experience and doesn't provide assurances that customer inquiries were actually resolved.

To gain a clearer understanding of the automated experience your company is powering, automated resolution (AR) should be your North Star and the goal of enterprise CX teams should be to improve AR over time.

AR can be more precisely defined as a conversation between the company and a customer that's relevant, accurate, safe, and doesn't involve a human agent.

Relevant means the AI effectively understands the customer's inquiry and provides information or assistance that's directly related.

Accurate means it provides correct, up-to-date information with respect to company knowledge and policy.

And safe means it interacts with the customer respectfully, not engaging in topics that would cause danger or harm.

The percentage of conversations resolved without escalating to a human agent can be expressed as automated resolution percentage (AR%), an important new key performance indicator for the customer service world.

To get a rough estimate of your current AR%, do the following

  1. Take a random sample size of your automated conversations, one that'll be of statistical significance).
  2. Manually tag conversations as Resolved, Not Resolved, or Unclear.
  3. Calculate the percentage of Resolved conversations.

AR and AR% are the cornerstones of measuring success in the AI-first CX era. Adopting these metrics and reporting on them accurately and efficiently should be your first priority. We've been developing and fine-tuning Ada's AI to be able to do this automatically and do it with more accuracy than a human.

Market competition across customer-facing verticals is increasingly steep, so maintaining close contact with customers is critical for continually improving products or services. Setting clear standards and parameters for customer support and using AI to gauge the automated service experience enables companies to keep the feedback loop closed and the business moving in the right direction.

Keep an open and active two-way channel between customer service and executive leadership. Customer service is poised to become one of the most valuable organizations in enterprises, and properly employing an AI-first CX strategy will ensure it rises in prominence. By enabling you to stay close to customers, you'll have a goldmine of data and insights that inform core company decisions, like what to build or market next, and this goldmine will only get bigger as you scale.

Mike Murchison is founder and CEO of Ada.