The future of customer service will be more automation through AI
If there’s one thing most customer service experts agree these days it’s that the future lies with more personal, empathetic service. Nobody wants another automate “Your call is very important to us” email. The shift to basic automation and outsourcing in the early 2000s allowed big brands to scale customer service quickly but it cost them on the quality of service. The industry got so bad that even now you can stand out from the competition by just not treating customers who have a problem as an annoyance.
The recent breed of startups like Transferwise, Moo and Hello Fresh have learned from these mistakes and taken inspiration from the big players, Apple and Amazon, to turn customer experience into a significant competitive advantage. It’s not going to be long until you simply cannot compete in any industry if you have the level of customer service commonplace 5 years ago.
So, how can companies scale their support functions to millions of customers while preserving and even growing the degree of personalisation and empathy? Ironically, the answer lies with more automation, a lot more automation.
Let me be absolutely clear that this is not to suggest more canned responses, speaking to annoying robot voices or knowledge bases full of irrelevant information.
This no longer needs to be the case, with the new iteration of AI. Companies can now build automated solutions which are actually as good as humans. Of course, we are nowhere close to being able to replace humans altogether, but if used carefully AI can dramatically boost the productivity of customer service teams.
Self-service on steroids
Self-service is the holy grail of customer support. Most customers actually prefer to be able to find the solution without talking to anyone. And for a service team, intelligent self-service could eliminate a lot of grunt work answering very basic, repetitive queries.
FAQs and knowledge bases have been around for a while and have definitely helped but do not go far enough. Modern AI’s can blend in the query, written in natural language, as well as any other customer attributes like number of logins, products purchased etc, to find a better solution.
Instead of using a very rough set of keywords, it can analyze data to find the most appropriate solution faster. It can also combine a lot more data sources under one roof, looking through official knowledge bases, customer support forums, Q & A websites and even social media.
Where keyword search would just provide the customer with a barrage of mostly irrelevant information, AI can find only an applicable solution in real time.
Giving your agents superpowers
Finding the relevant context is not just important for self-service. You agents will need it too. No matter how well your agent knows your product, they will not necessarily have come across all possible issues. Of course, some problems will be brand new, but others could be solved by just helping agents find an answer faster among the internal knowledge base. And at huge scale with rapid growth in the team, a new agent might not even know the product that well to begin with.
You can train the AI on all of the past support queries and after a few hundred thousand, you will have an AI that can suggest a solution from the archive more often than not. Your agents can reply to queries faster and with higher quality answers without harming the experience. They can then spend the extra time they saved on the really tricky problems. Once those are solved, the AI can become even better for the next time a problem like this comes along.
Personalization at scale
If there is one thing customers universally hate, it is when they perceive the solution is a generic reply, not tailored to their specific problem. Faceless generic replies such as “Your call is very important to us” are rightly ridiculed by the modern customer service teams. If my call is important, then why am I on hold?
With modern AI you canpersonalizee automated responses on the fly. This can be as simple as using the right language or as difficult as enriching the generic solution with specific data relevant to the support query. If done correctly, the customer will come away happy that you spent time understanding their specific problem.
Attitude at scale
It is a huge challenge to maintain consistent support culture across a team of 10 agents. It is a nearly impossible task when your team grows to a 100. What about 1,000 agents? Our analysis shows that perceived attitude problem from a customer service agent almost always kills customer’s overall experience even if the query was dealt with efficiently. This can cost you real money very quickly.
Modern AI can help agents moderate their tone and language in real time, allowing your customers to get consistent service 24/7.
A lot of the innovations described above are still beyond the reach of most companies today but an exciting new breed of startups including Chattermill are gradually making this possible. If you are building a best in class customer support function, you should already be using AI.