Are customer service agents in your team allocating their time wisely? Making sure that a support team has the right skills, tools and processes to be successful is becoming ever more tricky. On the flipside, if you understand and master the best practices and solutions available today, you could unlock incredible gains that will benefit your team’s effectiveness.
Contact center managers across industries mention maximising productivity as their key objective. Managers also care a lot about maintaining high quality; any initiative to increase productivity must ensure that they can continue to provide great service to their customers.
Increasing customer support productivity is therefore a difficult challenge to crack. Our mission is to help managers navigate these complexities by exploring the existing challenges and the available solutions.
Improving productivity – where to start? Before managers invest time and resources into implementing any solution, it is imperative to understand which problem is the most pressing – where exactly in the support agents’ workflows can you find the greatest opportunities for improvement?
We decided to diagnose the existing customer support agent workflow, aiming to detect and measure the potential problems and inefficiencies. We spoke with over 50 customer service agents, team leads, QA supervisors and operations managers across 12 companies of different sizes and industries. We also conducted time-and-motion studies with 30 agents.
Here is what we have discovered.
The chart above outlines how much of their time agents spend on each activity. Suppose an agent receives a query requesting a refund. Here is what they would do:
Read – understand that the customer is asking for a refund.
Information lookup – review the customer’s account information. Determine if a refund is eligible, and if yes, execute the payment.
Template search – search for and pick the relevant response template.
Typing – after the core template is added, add necessary content, such as refund amount, to make sure the response is comprehensive and personalised.
Editing – proofread, fix typos and improve the general language to the company standards.
Field filling – categorise the ticket as concerning a refund for future analysis.
Understanding the customer’s question As you think about improving productivity, note that the green stages in the chart above are specific to the product or service that the agent supports. For instance, a newspaper support team agent will look up information across different sources than a food delivery support team agent. Some teams may be dealing with specialised technical issues and require access to complex internal software, while others only respond to simple billing and account queries.
The information sources used vary drastically – these could be CRM/ticketing systems, external databases, company knowledge bases and others. All in all, the needed pieces of information are usually siloed across different apps, and siloed in different ways for each company. Agents need to have 3 – 5 different applications open at one time to fully address a ticket.
Are there any commonly used tools for improving efficiency across these green parts?
Given such diversity, there is no silver bullet that we have encountered yet. Each team must develop their own best way to improve productivity in these areas. There is one approach we recommend: depending on which external knowledge sources your team uses, you can often integrate them into your ticketing systems, e.g. via the Zendesk sidebar custom app framework. If successful, this could prevent a great deal of window switching and unnecessary data duplication, saving as much as 50% of your team’s precious time across the green stages.
One crucial thing that we observed is that agents know what they are going to write, before they even start writing. As such, there is significant potential for solutions that help agents express their thoughts much faster, than if they were to traditionally type them out on their own. A natural way to achieve that are templates.
Templates are indeed the first step that agents think about when composing a response: “do I have a template that I can use?”. Managers we spoke with tend to prioritise the development of a robust set of templates, typically ending up with 50 – 100 shared templates per team. To facilitate this, ticketing system developers have developed features to facilitate the use of templates – in the case of Zendesk, these are called “macros”.
The use of templates however come with their own challenges. First, the agent must be familiar with their team’s library of templates – it is not easy to memorise 50+ different use cases for which a template is available. As a consequence, agents very often end up typing out a reply, even though there are appropriate templates they could have used.
Additionally, for quality assurance purposes, templates are normally administered by a manager. It can take a significant amount of effort for an agent to successfully request any update.
What may surprise you, and what surprised many of the managers we spoke with, agents always end up composing their own lists of personal templates using text files or sticky notes. This can unfortunately lead to conflict between what is convenient for the agent, and what level of quality is acceptable to their managers.
Templates are a useful tool that on the whole improves productivity. There is however a point at which the benefit from creating templates diminishes. Customers can ask questions about hundreds of topics and it’s unfortunately not feasible to create, maintain and easily search through that many templates. It becomes a significant management overhead, templates get out of date and agents are not familiar with the less common templates.
Templates are the main composition tool only with a minority of tickets.
Most of the time agents end up typing many replies on their own, in “free flow”. In fact in 69% of tickets across companies we spoke with, no templates, whether shared or personal, were used at all. Responding to such tickets requires critical thinking, creativity and eloquence – domains where computer solutions have so far been notoriously bad!
All in all, typing is the single most time-consuming activity for an agent. Any tool that can significantly increase productivity during this stage will be key in maximising agent efficiency. In our research, the only tool we found that was widely used was Grammarly, but its impact was fairly low, as it typically helped only during editing.
There is nonetheless hope – right at this moment, a new generation of typing tools is being created, and it is showing a lot of promise. Tools such as TypeGenie are poised to change the way in which agents compose their replies, unlocking productivity gains previously thought infeasible.
There you have it: the issues and some of the top solutions used by email support agents. Making your team more productive is not an easy task. Here are some action items to get you started:
Develop a library of 50 – 100 templates – understand what the common responses are and create shared templates. The sweet spot between how many customer questions you can address versus how feasible it is to remember and learn such amount of templates seems to be around 50 – 100.
Invest in database – CRM integrations – if your team is large enough, you will benefit substantially from consolidating the information needed to compose a response inside the agent’s ticketing system. This includes external client databases and other relevant knowledge sources.
Use Grammarly – ensure that your agents’ replies do not contain embarrassing typos.
Consider innovative typing tools such as TypeGenie – as typing consumes the most of agents’ time, we have developed TypeGenie to fast forward the repetitive parts. TypeGenie already reduces typing time by 5 – 25% at many companies across different industries. You can try it for free right now!
This blog was brought to you by TypeGenie. TypeGenie is an auto-complete product for customer service agents. If you are looking to improve your customer service speed and quality, learn more about TypeGenie:Learn more about TypeGenie >>
A closer look at GPT-3 and the affect it may have on the future of business.
Github's Copilot offers a similar service to TypeGenie but focused on Developers rather than Customer Service Agents so we decided to take a closer look at it.