Even in this era of artificial intelligence and automation, customer service still depends on the human touch. Machines can handle common, low-complexity customer requests that are submitted online or made over the phone, but more complicated and nonstandard inquiries require swift and certain human intervention. Customer frustration builds when a reasonable request spins into a death spiral of handoffs from one agent to another as bad music plays in the background or when a chatbot delivers a string of stock answers.

Customer service is too important to get wrong, and yet many companies continue to optimize for cost and speed at the expense of the customer experience. They adopt an assembly line approach that is efficient for most simple requests but deficient for compound or complicated requests.

Encouragingly, leading firms across industries are experimenting with agile operating models in their customer-service organizations. The companies give multidisciplinary teams autonomy and accountability to solve any incoming request. The agile models have delivered spectacular results, but they are challenging to get right. Our experience working with many of these companies shows that the principles behind agile can be used to develop tailored operating models that deliver great service for customers.

Why Customer Service Often Falls Short

Setting up customer service as an assembly line—as many banks, insurers, telecom operators, and public-sector organizations often do—works well for typical requests such as changing a mailing address, checking a bank balance, and paying a bill, but not for nonstandard requests. Processes designed for efficiency become sluggish and error prone.

These problems are not new. In “The Agile Contact Center,” we diagnosed many of the reasons. For example:

• The overspecialization of agents, leading to calls and requests being handed off
• The use of legacy technology and siloed data, stymieing efforts to create a multichannel experience for customers
• The misalignment of incentives across the organization, inadvertently pushing customers to call centers when a problem (such as replacing a lost credit card) could be addressed more efficiently (for example, by making replacement a feature of a mobile app)

Automation Exacerbates These Challenges

As technology improves, companies are increasingly automating customer service requests. Their intention is to increase the speed and quality of service, while lowering the cost. Automation can be amazingly effective for routine tasks, but it is often incapable of dealing with the long tail of more idiosyncratic requests that require human intervention. When customer service agents don’t have the ability or authority to help finish the journey—or when they need to interact with several parts of the organization—customers understandably become frustrated. Leading firms are realizing that it’s time for a new approach—one that balances standardization, specialization, and responsiveness.

Agile in Customer Service

Customer service requirements vary depending on the industry, company, and context. A telecom operator with a simple selection of mobile plans must meet different customer demands than a competitor with a complex and configurable portfolio of plans. The customer-service operating model of each should reflect those differences.

At their core, agile customer-service operating models share the following set of principles that allow companies to swiftly satisfy customer needs:

• End-to-End Servicing. Teams have the skills to solve most requests. Handoffs and transfers to other teams are the rare exception.
• Autonomy and Accountability. Agents have the authority to act without running through a gauntlet of approvals, so long as they stay within an agreed-upon set of conditions. In turn, they are held accountable for the customer outcomes they deliver.
• Customer-Centricity. Customer service teams know and understand their customers and are recognized or rewarded for serving them successfully.
• Continuous Improvement. Teams and agents work on improving their skills and problem solving. Teams have feedback loops to improve performance.

To bring these principles to life, companies will need to address several key design questions to ensure that the chosen operating model is tailored for their specific circumstances. These questions include:

• How will teams be organized to solve customer problems? Will teams serve particular customer segments or needs?
• What stages of the customer life cycle will teams serve? Will the same team look after a customer’s journey from onboarding all the way through account closure?
• What level of training—generalist or specialist—is required to serve the customers?
• What support roles, such as coaches, are required to enable effective collaboration and productivity?
• How should teams leverage agile ceremonies, such as daily standups, retrospectives, and quarterly reviews?

In other words, companies must marry principles with practices to animate an agile operating model. When the marriage works, the results can be stunning. It does not matter if team members are remote, hybrid, or in close proximity or whether some of them are overseas or even whether many of them work for outsourcing partners. In all these settings, the performance of customer service can improve considerably: customer loyalty scores can jump by 10 to 30 points, repeat calls can decline by 20%, and complaints can drop by 50%. At the same time, employee satisfaction can improve by up to 70% while customer service operating expenses can fall by as much as 30%.

Agile Customer Service in Action

Three examples from different industries illustrate how agile principles and practices can improve customer service, but the operating models work across a wide variety of industries, including insurance. 

Energy. A fast-growing European energy supplier organizes customer service into teams of ten employees who are granted autonomy to solve customer problems. Each team looks after approximately 60,000 customers. If a customer contacts customer service more than once, the individual will almost certainly interact with someone on the same team and perhaps even with the same representative as before. The teams are measured on customer happiness and the number of customers that they serve. These metrics encourage a focus on both customers and cost effectiveness. The company has the lowest cost to serve in its home market.

The company has also built a system that integrates billing and customer relationship management. It manages all interactions, including calls, social media posts, and bills. Customers can use the intuitive self-service features to fix small problems, while others are quickly resolved through automation.

Banking. An Asia-Pacific bank wanted to create a more seamless experience for its small-business customers that apply for loans and other financial products. The bank created teams of about 15 people with broad skill sets. The teams were responsible for all customer requests from a specific region, and six or seven teams reported to a single manager. The bank gave these teams both full autonomy and accountability, and they responded by improving the employee experience and business results.

The teams managed their own daily scheduling, including shifts, vacations, phone coverage, and back-office work. The teams also had control over their workflow and learning initiatives. They identified bottlenecks in their processes and the additional skills that they needed to better serve the customer. The teams then developed these skills through self-directed, over-the-shoulder learning. This cultural shift was supported by extensive coaching to embed new behaviors. While there were initially concerns that stalwarts of the old model would resist, employee satisfaction jumped, and teams provided better, more efficient service.

Telecommunications and Media. A global telecom and media company wanted to use a big product launch to test a new approach to customer service. The approach was grounded in a culture of delivering exceptional customer and employee experiences through simplicity, honesty, and continuous improvement.

The company started by running several design sprints to define the new agile operating model. It then launched a team of teams with about 150 people who were responsible for servicing all the requests of new and existing customers; these teams were made up of agents who were generalists and specialists, team leads, and coaches. A change team was responsible for gathering feedback from agents and customers and analyzing the data to improve a service or lower the cost to serve.

Most agents had multiple skills, and they were able to solve many different customer queries end to end. If not, though, agents would ask for help within their team and learn from the experience so they could solve similar issues in the future.

The early pilots suggest that this way of working will likely succeed in increasing customer loyalty scores, call efficiency, and employee engagement scores, while lowering absenteeism. The company will now explore where else in the customer service organization this approach may be effective.

What Can Go Wrong

While agile operating models are easy to grasp conceptually, they can be hard to implement in practice. We usually see several common challenges.

For some companies, the model breaks down at scale. Companies should not declare success after one or a few teams implement a new customer-service model. Agile models are often easy to make work in pilots but hard to scale up effectively. Companies need to rethink their workflows to make sure that end-to-end accountability stays within a team as much as possible, yet allow for overflows to be distributed seamlessly to other teams where required.

In some situations, the productivity of teams using a new model can initially dip before it starts to surpass the productivity achieved using the original model. Implementing any new operating model is hard, especially when it significantly changes the way people work. Depending on the starting position and the amount of change, it can take a few months before teams get used to their new autonomy and start delivering markedly better service in a more effective way. This transition period may be discouraging, but it is common and should not be an excuse to give up early.

Finally, we sometimes see managers reverting to their old ways of working. Managers in customer service often have grown up on the operations side of the organization, and they are often comfortable with a command-and-control style. If productivity does dip initially or other problems arise, they may regress to what is familiar. Coaching can help these managers recognize that they need to let their teams become the problem solvers.

How to Start

To embed the type of change that is required by agile ways of working, it must be integrated across all areas of the operating model. Six tips can help put companies on a strong path.

Start with the why. Leaders must define and agree on why they want to start on a customer service transformation as well as what they expect to achieve. That way, the operating model can be designed to target the organization’s particular pain points and ideal benefits.

Tailor the customer service approach to the business. While the principles of agile customer service are universal, all companies and contexts are different. Organizations cannot simply copy a successful approach from another company.

Conduct pilots and experiment. The best way to design an operating model is by working with the customer service agents and by iterative testing and improvements before a large-scale rollout.

Choose the right places to target. Pick areas with high levels of handovers, rework, and complaints. These areas will likely have the highest return.

Focus on the culture and mindset, not just structure. The structure of the new operating model is important, but the structure alone will not entice people to work in new ways. Leaders and frontline employees alike need training and coaching in how they should be leading and working differently.

Minimize the disruption to customers. Introduce the new model through pilots and phases so that the transition is smooth. “Big bang” launches may be splashy, but they risk disrupting service for customers, especially if the team’s productivity temporarily dips while the model is becoming embedded.

If customer service was easy to get right, more companies would be getting the job done. But for most companies, customer service is still a work in progress. Companies can accelerate their progress by adopting a golden rule approach: treat customers as you would like to be treated.

The authors would like to thank their colleagues Nicholas Clark, Pim Hilbers, Hrvoje Jenkač, Sukand Ramachandran, and Michael Urban for their invaluable contributions to the creation of this article.