Robots are great for cleaning the floor and are perfect for exploring the moon. They’re just not that great at customer support. The last thing your customers want is another “We received your message” email or “Thank you for holding” recording. Robots only succeed in making customers feel like another number, a dubious accomplishment for your team. They’re the opposite of the personal touch that effective support is supposed to be all about.
It’s not that robots are useless. They’re great at repetitive tasks, perfect for finding data and remembering anything you’ve ever written down. As sidekicks, robots can help offer more personalized support, doing the tedious parts of support so that you can focus on actually solving problems. You just have to give them the right job.
Here’s how to find the perfect job for your robots, so that you can automate support and offer more personalized, hands-on support at the same time. We speak from our experience of working at Zapier, a little tool we built to make it easy to automate tasks between web apps.
Customer support, at its simplest, is easy. Answer questions, solve problems and make people happy — that’s all there is to it.
It’s not that answering questions and solving problems is so hard. It’s that there are so many questions to answer and so many problems to solve; so, getting to the end of the queue is hard. If you have to look up customer data and figure out how to fix each problem by hand, you’ll never have time to answer everyone.
That’s why it’s so easy to turn to basic automation for the appearance of faster customer support. If you annoy customers with delays, maybe they’ll give up and not ask you their question. At least that’s one less thing to answer, right?
Yet, that’s exactly the opposite of what support should be. You should value each question because it’s the perfect opportunity to give a personal face to your company and to make your customer’s day better. What better way to make sure people tell others about your product than to make sure they love it?
Sensible automation helps you do just that. Leverage robots for their ideal jobs, and let them find data, diagnose problems and answer the simplest questions automatically. Then, you’ll have more time to add a personal touch to the rest of your emails.
Filter Tickets To Reply Smarter
Filters aren’t just a handy way to see how many emails come in about certain topics or to search through emails from customers who have purchased a particular product. They’re also your best — and simplest — way to enlist robots for your support team.
Say you’ve answered one question about pricing, and you’re ready to answer another ticket. Answering another question about pricing next would be fastest because that topic is already on your mind. Instead of switching gears for the next ticket, use a filter to find related tickets, and reply to them all in a row.
You can also use filters to find the most urgent tickets, those that got left behind yesterday or those from people whose issue hasn’t been resolved in the last few replies. And you can use filters to get rid of junk messages and auto-replies from customers in a few clicks.
In many customer support apps, filters can do some support work for you as well. They can add tags to emails based on content, automatically assign tickets to the correct team member and more. As MistoBox co-founder Connor Riley says about Desk, their customer support tool, “You can do just about anything with Desk’s rules and filters without a line of code.”
Or you can make your own work easier by asking customers for more information up front. You could just have a simple
firstname.lastname@example.org email address for support questions — that’s what most companies use. Or you could have a detailed support form that asks detailed questions about users’ problems to help you pinpoint issues. With more data to go on, the filters can automatically sort messages much more accurately.
With just a few minutes of setup time, you can easily shave down your response time and make sure the most urgent problems are solved first, with less time spent trying to figure out exactly what the issue is.
Use Macros To Reply Faster
Pre-written auto-replies are the easiest — and laziest — way to employ robots for support. Don’t use them.
Do, however, use pre-written emails to send personalized answers more quickly. Say you often get similar questions about your pricing plans. Write up a standardized reply that fits most scenarios, and add it as a macro to your support app (or use a tool like TextExpander to add it with a keyboard shortcut).
But don’t stop there. Go ahead and add something personal to the email — perhaps greet the customer by name, answer anything else they’ve asked, and then tweak the pre-written section about pricing to sound like the rest of the email. Your customer will get a personalized message, while you’ll spend a fraction of the time answering recurring questions.
If you get the same questions often enough — something you can quickly figure out with macros, too — that might be a good indicator to update your knowledge base with more detailed documentation. You might still get questions about that topic, but that will at least cut down the volume a bit. Watch indicators like that for things you need to improve in your support — and keep an eye on support metrics such as response time and customer happiness to keep refining your support replies and improving your customer support over time.
Solve Simple Problems Automatically
Now, for some problems, an automated email just might be the perfect solution — even better than a personalized email message.
The HubSpot team, for example, noticed that many of the support emails from new users were about their confirmation emails. “Inevitably, some people do not receive this email,” says HubSpot product manager Sam Awezec. “Any support team would strive to get these resolved as quickly as possible, but at best there would still be a 20- to 30-minute delay.”
Time to hire a robot. They added a filter to watch their customer support app for messages that mention “setup” or “confirmation.” Then, they used an app integration tool to check HubSpot for whether the user’s account was pending. If it was, Zapier triggered a HubSpot workflow to send the user a new confirmation email.
Crisis averted. The customer would receive another confirmation email within five minutes, and the support team could spend their time fixing problems that aren’t best handled by a robot.
Go through your most common customer support messages to see whether any common steps could be effectively solved with an automated workflow or message — the ones that could be solved just as effectively with an automated reply as by your team. If there are any, then perhaps use auto-replies and some automation to solve those particular issues — and then use a personal touch to solve the remaining problems.
Automated emails are good for things that robots do best: extending trials, confirming appointments, sending license keys and other things for which you would merely have to look up and copy data. A robot can reply faster than a human, and with the exact same data. For real, larger problems, though, individualized replies are best — you will be able to address those issues personally and make sure your customers are happy.
Surface Customer Data For The Full Picture
Solving those trickier issues usually takes more time. You won’t be able to rely on a pre-written message, and an auto-reply from a robot would only make things worse. Instead, you need to dig in, figure out the solution, and explain it to the customer.
Robots might not be the best at replying, but they can help you find the correct data. You can use automations — ones built into your customer support tool or that come from third-party integration apps — to pull in data about each customer. Perhaps you store information about customers in a CRM and have a payment-processing tool; integrate those with your support tool and you’ll be able to see personal information and perhaps the product a customer has purchased. Sending a refund or calling the customer, then, will be only a click away.
If you pull as much data into your tickets as possible, your team will spend less time searching for information and more time helping customers. For instance, we’ve added custom integrations to our Help Scout account; the sidebar shows internal documentation about the apps a customer is using, along with quick links to their account and more. The Mistobox team uses Mixpanel webhooks to create cases and add customer fields in Desk; then, they prioritize tickets based on that information. “It allows us to get back to our most valuable or at-risk customers faster,” says Mistobox’s Riley.
These types of customizations will take more time, but they are the most effective way to help your team solve problems for each customer as quickly as possible.
Automation will also bring all of your tickets into one inbox. Don’t check Twitter, Facebook and a half-dozen email accounts each day; instead, connect them all to the same help desk, so that your “New Ticket” count shows everything you need to answer. You can even have customers help you help them by filling in a support form in which they describe their problems in detail — then, you can connect that form to your help desk with integrations, and the information will show up with the rest of your tickets.
Another trick is to use chatbots to look up information from your help desk right inside Slack or another team chat tool. Your support software may have its own integration with your chat tool; otherwise, you could build your own chatbot quickly using a customizable integration tool. Just have the integration watch your chat conversation for keywords, like “/email Bob Smith”; then, set it to search your help desk for emails from “Bob Smith” and send links to those messages back to your chat app. That’s a quick way to look up information without having to open yet another app.
Automatically Translate Tickets
Putting together a support team that knows all languages would be impossible. Each person on your team should be able to answer any question that comes in — and for that, you need translation.
Google Translate is the quickest way to get anything translated into a language you can read, even if it’s not the most accurate. Google Chrome will even offer to automatically translate support tickets and other pages for you, without an extra click. Or use the Google Translate API to automatically translate support tickets, for $20 per million characters.
You have other options, too. Unbabel, for instance, offers professional translation from real human translators. It’s the perfect mix of robots and people: It combines Zendesk integration and an API to automatically send tickets from your app to Unbabel, along with real human translators for perfectly translated messages. That’s why the Trello team uses it to translate their tickets; the translated messages are pulled back into their help desk automatically.
Build New Support Tools
Sometimes you’ll need something more to help out customers, something that seems to require an entire development team and a new product. That’s tough to get, and it’s easier to just manage without it.
But there’s no reason you can’t build your own support tools, even if you’re not a developer. With a form, a survey database tool, a spreadsheet and some integrations, you can build an entire mini app just by mashing up existing tools.
That’s how Disqus’ support team built their first abuse- and spam-reporting tools. “We knew we wanted to use a survey format in order to get all the abuse report information we needed in order to investigate, but we also wanted to use Desk to process the completed surveys,” says Disqus product support manager Daniel Matteson. So, the team turned to their existing Wufoo account, made a form to gather the abuse data they needed, and used Zapier integration to send that to Desk.
That worked well enough to route the first several hundred reports into their support center and to enable them to work around the problem without any time spent on development. Then, when Disqus’ development team had time to build a better reporting tool, the form served as the first mockup, which helped them build the final version even faster.
Perhaps you need a more complex tool and think a basic non-coded solution wouldn’t be enough. You might be right. But today’s robots are pretty clever, and plenty of off-the-shelf apps would likely solve at least half of your problems with minimal work. Patch them together, and you just might have the perfect solution — or at least something that will get you by for now.
Giving Human Support
Automation won’t solve everything. It might help you find information faster or answer the simplest support tickets without typing, but it still doesn’t have that human touch. For that, you’ll need to answer questions and solve problems yourself.
Sometimes that’s easy — but often it’s tough to reply to the same questions over and over again, putting out little fires as quickly as the pop up. Here are some tips on handling the toughest support challenges that automation can’t solve:
- Listen. People get upset for a wide variety of reasons, and you can’t brush off their issues just because they’re angry. Instead, you need to listen to their whole story, empathize with it, apologize for the trouble, try to resolve the issue, and then diagnose the core problem with your team so that it doesn’t come up again. That’s the HEARD technique, something Disney uses to improve its amusement parks.
- Say sorry. If there’s one thing we can do far better than robots, it’s show emotion. Sometimes things will break and go wrong, and all you can do is say sorry. Do that. Explain the issue, be willing to say you’ve messed up, and more often than not people will understand and not blame you. And if they do blame you, don’t take it personally — just use the incident to improve your service going forward.
- Be positive. You can’t offer a discount to everyone or build every feature that’s requested. But you can listen to what people want, and word your replies positively to help them focus on what you do offer.
- Be honest. You don’t know everything, and that’s OK. Don’t try to weasel your way out of tough questions. Be willing to say you don’t know stuff sometimes — people respect honesty.
- Respond. Maybe you can’t fix an issue right away, or perhaps there are too many questions to reply to everyone today. At least send a simple — personal, not robotic — message acknowledging the issue and giving the customer a timeframe for when they can expect a reply. That’ll let them know they’re not forgotten and will give your team some breathing room.
Building Your Own Effective Support Automation
Automated replies that actually solve problems; translation on the fly; new tools built without a single line of code — that’s smart support automation, and it’s not that difficult to start using with your team. Just break down your support challenges into steps, and figure out where pre-written text and clever automation can make things simpler.
Then, don’t sit back. Use that extra time to offer more personalized support, and you’ll have happier customers — ones who are far more likely to be fans of your product for life.
Looking for the perfect app to manage your support messages, ideas for building an effect knowledge base, or tips from Trello, Intercom and other startups on how to support your customers? The free Ultimate Guide to Customer Support eBook — which this article was adapted from (chapter 8, to be precise)— includes that and more. Read it online or download a free copy for quick tips to supercharge your customer support.
Or check out these other great articles about customer support:
- “Idiots, Drama Queens and Scammers: Improving Customer Service With UX,” Laura Klein
- “Supporting Your Product: How to Provide Technical Support,” Rachel Andrew
- How To Deliver Exceptional Client Service
- Improving Customer Service with UX
- Taking A Customer From Like To Love
- All You Need To Know About Customer Journey Mapping