Automated customer support — a technology that can automatically resolve certain issues for users, is widely used across industries and businesses. Its purpose is to reduce the rising costs and demands on support agents in helping users, which becomes increasingly needed when a company is scaling. 

Of course, that’s not to say you can leave automation to its own devices, unchecked and unmonitored. Some processes or interactions need a human behind the screen to provide that distinct human touch. But other actions can and should be left to the efficiency and speed of automation. As they say, it’s all about balance. 

For mobile apps, customer support is required both in-app and on the stores for user reviews. With that in mind, the best setup for your support team is the combination of human agents and automation.

Examples of customer support for apps 

But what solutions can an app business implement to make automation have the most impact and enable support agents to work with ultimate efficiency?

Automation solutions

In-app support Support for app stores 
– Self service 
– In-app messaging and chatbots
– Interactive FAQs
– Help bot automation 
– Asynchronous messaging 
– Intent classification
– Issue routing
– API calls
– Auto-replies & auto-tags 
– Auto-report & auto-translate
– Review filtering and segmentation 
– Smart alerts and reports
– Templates library (pre-made rules)

Solutions for the agent 

In-appUser reviews 
– In-app chat
– Agent desktop
– Quick Replies templates
– Skill-Based Routing
– Untethered messaging
– Customer service analytics
– Response templates (check out our template guide for some examples)
– Semantic tags
Semantic analysis 
– Ratings & Reviews dashboard 
– Reply to reviews 
– Agent performance
– Compare feedback 

When to use customer support automation

So now you have a good idea of the app customer support tools for both agent and automation. But what tasks do your agents take on, and what can rules-based automation resolve? The reality is that agents and automation excel at different tasks. 

There are those in the customer support industry with concerns about the overuse of automation resulting in companies losing their human touch or support becoming too robotic. It’s important to note that automation doesn’t have to be robotic or impersonal. The technology has gotten to a point where chatbots or review responses, for example, can interact with customers based on their use case and phrases used. Contextual routing goes further by taking into account the user’s account details and their particular issues to find them the best solution before they even speak to an agent. 

The outcome is that the customer feels acknowledged and seen — thus, is more likely to continue the interaction to find what they’re looking for. Automation which involves contextual routing is especially useful since, according to Salesforce, 79% of customers were more willing to share their information when they felt understood through contextual interaction.  

Your customer isn’t communicating with your system looking for a friendly chat — they have a problem they solved quickly and accurately. Many users with issues can be routed by a bot to the right solution straight away. Alternatively, their problem and information can be tagged and categorized in the dashboard for an agent to retrieve later, rather than keeping the customer waiting in a queue and asking for that same information all over again. Having to repeat details was discovered by Hubspot to be one of the most frustrating aspects of customer support for customers. Contextual routing, then, would be an improvement on how issues are handled and reduce the steps to resolution, which may have a positive impact on CSAT. 

So, where does automation work? 

A classic situation for automation is with simple user queries that come through in-app messages or reviews. There’s no need for these to sit in a support agent’s ticket pile when they could be resolved by a bot in the blink of an eye. Think of it this way: something that can be fixed in less than a minute should not wait an hour in a ticket queue.

Of course, there are many other situations in which it would make sense to implement automation, such as:

1. As your business scales, your support team is struggling to catch up with capabilities and resources.

2. When your app has a global user base, you need to offer multilingual support but don’t have a multilingual support agent.

3. To provide support when it’s out of office hours for your agents

4. Maintaining query classification routing and mass spam reporting.

5. When your goal as a support team is to cut down response time or increase the response rate for your in-app chat or app reviews. 

6. Automatically categorizing customer insights for other teams. 

7. Cancellation for those users who are 100% sure they want to terminate their contracts.

What to prioritize for the support team

However, it doesn’t make sense to use customer support automation for everything. Humans have interpersonal skills and empathy, which are necessary qualities for complicated issues that sometimes involve emotion. The situations that agents should manage are issues like customers wrongly charged, a complaint on the experience, queries about the business or something beyond the product experience itself. If it’s important to the customer, about which they feel emotional and require some empathy, they’ll probably be willing to wait for a support agent. 

When it comes to complex issues, customers will be seeking out the experience of a support agent. Being unable to do so could turn them against the brand. In one report, 30% of those studied said that the most frustrating part of customer service is not being able to reach a human agent. Not having a good support structure in place, where tasks and tickets are suitably split according to issue type between agents and automation, can turn off customers and lower your CSAT. 

Another key prioritization for the support agent is high-paying customers. You do not want those paying the most to be waiting in long queues to get the assistance they need or leave their review without a response for days. At the same time, your agents shouldn’t be solving issues for every user or responding to every review. It’s a matter of dividing up who gets what type of assistance based partly on monetary value. If you’re a customer that’s spending a lot on a product or service, you’d expect nothing less than personalized, human support. Low paying or free users can be sent through self-service and can only reach an agent when they need help with a complex issue.  

Obviously, customer experience needs to be consistent for all customers, and they always should receive the help they need. But with finite resources, you have to prioritize agent time to those with the highest value to the business. 

So, the type of things best suited to a human agent is:

1. Complex issues that need the interpersonal skills of a human.

2. When emotions are running high. The authenticity of human empathy and sympathy is required here.  

3. Unusual cases not covered in help articles, which can’t be resolved through issue routing or response template either. 

4. Decision making. The actual analysis of the support team’s results and adjusting strategy accordingly based on the data. 

5.  Onboarding new customers. This could not be done by automation alone.

6. Prioritizing personalized help for your highest paying customers.

Automation is there to help, not replace the human agent 

The most significant takeaway here is to see automation as an opportunity to improve the productivity and efficiency of your organization rather than something that will radically or drastically change the very existence of your support team. At the same time, anyone with the opinion that automation can do everything for mobile support without the need for human agents is just wrong. Bots cannot empathize like a human, nor do they have the cognitive flexibility of a technically specialized support agent. 

But, in the age where a consistent and prompt customer experience is a determinant of future growth, having all the skills and tools at your disposal is simply sensible. This includes:

  • A well-trained support team 
  • A comprehensive, well-written knowledge base 
  • Rules-based automation for in-app support and review management 

It’s important to remember that there are no right or wrong tools, automation, or systems. As Flo Health said in their case study with AppFollow: “there is no incorrect automation — only poor set rules and triggers.” 

So what does the ideal outcome of correctly set automation look like? 

CSAT is a key indicator of how happy customers are with the product or service. There are some understandable concerns about automation negatively impacting the CSAT score. But Helpshift’s Player Benchmark Report found this was not the case — and, in fact, it shows that the reality of automation’s impact on CSAT is quite different. 

This image from the report shows that when Helpshift’s clients combined the use of chatbots with support from their agents through in-app messaging, it resulted in the highest CSAT of all support setups, at 4.3 — higher than those fully manual (4.2) or fully automated (4.0). 

The report also demonstrates that when companies increased their use of automation, they saw higher CSATs and no evidence of the increased use of automation damaging the support experience for their customers.

As you can see from the chart above, the frequency of issues solved through automation or partially through automation by Helpshift’s clients increased substantially and steadily over two years. There was a 46% increase in the use of automation between February 2020 and May 2021. In terms of CSAT, customers saw an average of 4.1 – which is a high level of performance – and two-thirds saw their CSAT increase. 

This proves that using automation does not wreck CSAT if used correctly alongside human support. It can actually have the opposite effect. 

Why is this so? For two reasons:

  • Helpshift’s customers who used automation saw a 60% decrease in Time To Resolution, meaning that issues were resolved much faster than previously. This improved efficiency satisfied users.
  • Support agents could focus more on resolving complex issues as automation handled the simple, repetitive tasks. This made users feel cared about, increasing their satisfaction with the service.  

A good example of using review management automation

Bitmango is a hugely successful global mobile game publisher with over 10 years experience in the market. They have a large accumulated user base — with over 472 million daily users and 710 million cumulative downloads. Across the 100s of games they support, hundreds of user reviews flow into the app stores every day. 

The issue was that Bitmango could not manually respond to each and every one of these reviews without overwhelming their support team. Agents were spending time replying to repetitive reviews that included simple queries and leaving ‘thank you’ notes instead of focusing on reviews that required the personal touch. 

So they turned to AppFollow and implemented auto-reply to help them manage the reviews coming in. The idea was to use it to tackle positive reviews and simple complaints. They set 10 rules-based automation for review responses with 15 different response templates for each typical issue.

In three days, their speed of processing reviews increased by 2.3 times and they were able to cover 2000 reviews. They also saw:

– A huge rise in their response rate from just under 40% to over 80%, as a result of auto reply.

Graph 1: the impact of auto-reply on average rating & response rate

– A reduction in response time for reviews — from 72 hours to just 24 hours

Graph 2: average response time of auto-replies vs without auto-replies

These results prove several things. Firstly, it shows that automation will not seem robotic to users and ruin the customer experience when used right. AppFollow’s randomization algorithm provides a variation of responses to user reviews, which will prevent everyone from being sent the same message. Bitmango’s use of it shows no negative impact on their average rating, as presented in graph 1. Secondly, by splitting reviews between automation and agents, the support team was able to successfully deal with the reviews for which a human does best— updated, emotive, and featured reviews. Finally, when implemented and used following best practices, automation works. It improves a number of key metrics for a support team, including response rate, response time, CSAT, and average rating. 

A good example of using in-app support automation

PagueVeloz is a Brazilian pioneer in financial technology which harnessed the power of Helpshift’s automation to improve its customer support processes. The company was finding it hard to upkeep internal operational processes and customer support that was still largely done manually. It was inefficient and slow, as a result. 

After implementing Helpshift’s chatbots, PagueVeloz was delighted to see: 

  • A 50% decrease in ticket resolution time
  • An increase of 13% in CSAT
  • Handling 173% spike in tickets without hiring more agents

This shows that when automation is implemented it actually has a positive effect on CSAT. This was mainly because issues were resolved a lot quicker than previously, which meant fewer frustrated users. 
In the words of Guilherme Bulhmann, Head of Customer Success at PagueVoz:

“The bots are awesome. How you use them, how you can build them, and how they can respond to your company’s needs. You can create some really great things with just some clicks. It’s incredible… We use all of it because it helps when a customer needs to chat.”

Conclusion 

So what does this tell us about customer support automation for apps? 

It’s hugely beneficial to your bottom line, helping to cut costs and improve support efficiency. This is because it helps the support team agents to nurture customer relationships and  focus on the complex issues, while automation covers the repetitive, simpler tasks. Customer support automation in some form is necessary as your business scales to ensure you continue to provide a prompt and effective customer experience. The make up of a solid customer support strategy involves both automation and support agents. They should complement one another to do what they do best in keeping users satisfied. 

Find this balance with your in-app support through the Helpshift platform — Book a demo today.

Or, to find out how you can master review management automation, visit AppFollow

Sasha Hodes
Content Marketing Manager
AppFollow 

Published May 24, 2022
block background image