So, how does a company scale up, grow fast and still deliver world class customer service?
If you ask our customer Gousto (contact centre of the year 2021 and 2020), the secret is automation.
Gousto automatically handles around 40% of contact volume using automated workflows, artificial intelligence and machine learning.
They also use AI for automated ticket tagging.
As the founder, Timo Boldt, says, Gousto is a "data company that loves food".
In this article, through research and expert interviews, we'll explore how you too can:
All using AI-powered tools.
There are three categories: tagging, self-service, and productivity. Each of which bring a number of benefits to you as you scale.
Before you continue, this is article is part two in our scale up series. In part one, we cover six non-AI solutions for scaling customer support. The articles make sense independently of each other but we still recommend starting there.
AI technology is radically redefining the customer service landscape. However, in its current technological state, AI is GREAT at some things but terrible at others.
So what is AI good at?
Paradoxically, computers are really good at doing things that we find hard, like finding the square root of 3485769.92. But struggle with things we find easy (like lifting a robotic arm).
In the context of customer service, it’s difficult for humans to read 10,000 support chats and accurately categorise them into specific issues or reasons for contact. Whereas, the latest developments in natural language processing AI can do that in seconds.
On the other hand, AI still fails miserably at talking to customers (now you know why chatbots have a bad rep!)
When looking to AI as a solution to customer service challenges, keep in mind that AI is currently great at:
In the following section, you'll see these 'abilities' in action.
Goal: Power automations and extract valuable insight to drive company-wide improvement.
With that ability, AI software can accurately tag every support ticket at a granular level. SentiSum's AI, for example, will tag tickets with multiple topics, sentiment, priority level and more. So, rather than a generic tag like 'complaint', you'll know exactly what's driving the complaint.
AI tags can also be used to power time-saving automations that are critical as you scale. Since the tags are applied in real-time, they can be set up to trigger automated workflows like auto-triaging that ticket to the right team or auto-prioritising a ticket based on its contents.
If we zoom out to the macro level, AI-based tags allow you know exactly what's driving customer contact.
In other words, you’ll now have the ability to create a clear path toward tackling the root cause issues.
In an ideal world, customers would never need to contact customer service. The best customer service is no customer service.
So when they do contact you, it’s important to listen, synthesise that feedback, iterate and improve your products and internal processes.
An accurate AI tagging system will ensure support insights are listened to company-wide.
We recently wrote an article detailed all the use cases of AI-based ticket tags.
For this article, we'll narrow it down to those that help you scale.
Our customers continuously feed their product teams with AI-powered tag insights. Armed with those insights, they can implement changes that tackle the issues customers are facing.
We're not just talking about technology products here.
For example, if you’re working for an eCommerce company, support ticket insights could:
If you'd like to solve operational issues—like delivery or packing issues—AI tags help you identify the root cause issue quickly and, therefore, reduce time-to-fix.
Each of these use cases leverages the power of frontline insight to drive innovation and development.
If you'd like to focus on internal efficiency in customer service, support ticket insights help there, too:
It’s important to remember that for every customer who contacts support there is 40 to 60 more than didn’t bother to.
If one customer wants to buy from you, but can’t because a bug in the payment process stopped them, there could be 60 more potential customers who never became a sale.
As you grow fast and scale, all of the support leaders we talked to identified the importance of understanding exactly why customers are contacting you.
As a support leader, especially as you scale, it’s not uncommon for the CEO or company founders to ask you, “why are our customer contacts increasing so much?”
It’s also common for product managers to ask, “what are customers saying? What can we do to improve things for them?” or “we have a new project starting around ‘X’, are there any support insights we can use?”
When these scenarios arise, it’s important to have an answer at your fingertips.
With an AI-based tagging system in place, you’ll easily be able to answer questions like these, understand the root causes of issues and reduce the time it takes to see a fix happen.
Goal: Reduce contact volume by dramatically improving your self-service
AI is brilliant at automating repetitive tasks.
When it comes to self-service, the primary use of AI is to read and understand human language at scale and trigger an appropriate response to it.
Prior resolved dialogues with customers serve as training sets for each of the use cases below. So your historical support ticket and the way agents have handled queries underpin how an AI helps customers in the future.
Here are four ways AI helps with self-service:
Often, your customer community or existing knowledge base already contain the answer a customer is looking for.
The only problem? They can’t find it.
An AI tool like Yext can change that. Yext builds a powerful search engine for your website.
Most websites have a simple keyword search, surfacing every article that contains the word searched for. With a tool like Yext's AI, you'll have a more Google-esque search experience.
We interviewed Graham Johnston, Head of Omnichannel at Three who had just implemented Yext:
"It allows you to have really good control of what customers get when they search for a particular question or a particular phrase on the site search, rather than just getting a list of blue links."
"You get a Google-like response with tabs and all of those searches can be customized. So we can customize what customers see as a result of the types of searches that are being made and spin up very easy FAQ's and content to help customers to achieve what they want to achieve digitally."
Sounds great, doesn't it?
Oh, chatbots. We’ve all heard of them, we’ve all hated them.
AI chatbots are only really capable of answering routine and basic customer questions.
The AI recognises a question, probably based on the presence of a keyword, and automatically replies with a templated response or article from the help centre.
We suggest making a chatbot a very frontline of customer service. But, to maintain high customer satisfaction, allow customers the option to ‘talk to a human’ from the get-go.
An intelligent approach to designing a knowledge base looks at the overlap of easy-to-answer questions and high-enough-volume-to-be-worth-it questions.
With the AI tagging system in place, you'll have why customers are contacting you at your fingertips. That insight makes it easy to identify where there are opportunities to add additional help centre articles.
By covering more and more bases with help centre articles, you can have fewer agents and therefore scale more efficiently.
We've seen that AI can read support requests in real-time and tag them with why customers are contacting you.
With that system in place—where a computer understands the topic of conversation—you can easily leverage it to send templated answers and relevant help centre articles.
This method will close more contacts on first resolution and reduce the number of queries handled by human agents.
Goal: Speed up agent response time (and make them happier!)
AI can augment the way agents respond to each customer request, helping them be more productive at handling tickets.
AI can help keep your headcount low as you scale by handing back time to agents on every single ticket.
Over 95% of customer service responses are repetitive and therefore can be predicted.
With the integration of a sentence completion AI tool like TypeGenie, your agents simply begin writing and have suggested answers appear.
Over thousands of conversations, sentence completion adds up to significant improvements in handle time
In a similar way to sentence completion, AI can help save agents time by suggesting entire template responses.
Some queries are so common that you can simply copy & paste an answer to them.
With an integrated tool, AI could suggest a macro and that could be a one-click response for an agent.
At the beginning, we said that the scaling support team faces three fundamental questions:
We've showed you how AI moves the needle on all three in a way that's cost efficient.
Improvements in CX and satisfaction stem from faster response times, as well as support-product feedback loops.
Whereas, AI helps the scaling team remain cost effective by support the reduction of ticket volume.
Finally, where AI comes into its own, helping the entire company grow and innovative. AI-based ticket tagging helps extract useful insights that empower change and improvement company-wide.
AI is being used in customer service to help customers get solutions faster. Via on-site search and agent productivity, customer support teams are able to respond quicker or deflect tickets to help centre articles. AI is also being used to extract insights from support conversations. With the latest developments in NLP, AI ticket tagging can provide an accurate and detailed understanding of what friction and pain points customers are experiencing.
Artificial intelligence has become the future of customer service due to improvements it provides for both customer and company. AI ticket tagging, in other words extracting customer insight, in particular is being used to underpin strategic change within organisations.
AI helps in improve client satisfaction in two ways. Firstly, AI helps businesses extract insights and understand what issues are driving customer contact. With those insights companies can identify improvement areas quickly and improve the customer's experience. Secondly, AI speeds up the response time by answer queries automatically. Lower response times mean less customers sat in a queue waiting for help.
AI is changing the customer experience in two ways. Firstly, AI helps business understand what's driving customer contact by extracting granular insights on the topic and sentiment of each conversation. Secondly, AI speeds up time-to-first-response, ensuring customers are responded too quickly with helpful articles and template responses.
AI is often hit or miss. If you're looking for accuracy and efficacy, it's important to test AI on your own data. When you see it in action, only then will you know it's safe to buy.
Machine learning is a subset of AI, it allows AI to answer and understand language in complex environments. Machine learning is being used in customer service to extract high-quality, granular insights, improve self-service, and power chatbots, as we've deep dived earlier in the article.