On daily basis, we hear one thing new about AI – what it’s, the way it’s evolving, and the way it will affect us.
There’s nonetheless a lot we don’t find out about this expertise, however there are some myths we are able to already bust – at the least in terms of customer support. As time goes on, and rumors unfold, it seems that similar to a nasty AI chatbot, people are fairly good at simply making issues up.
So, listed below are among the myths we’re seeing about AI in customer support.
AI will take all customer support jobs
It is a widespread worry throughout a number of industries, but it surely’s already changing into clear that distinctive human strengths will proceed to play a significant position in top-tier buyer help. Whereas AI takes on the extra repetitive duties, help groups have an opportunity to suppose deeper about their buyer expertise providing, and handle the more complex cases that in the end enhance buyer loyalty.
As Des Traynor, Intercom’s Co-founder and Chief Technique Officer, stated on a recent Intercom podcast: “Excessive urgency, excessive drama, excessive emotion. These are the varieties of issues the place people specialize.”
“The much less time spent on repetitive FAQs, the extra time reps can spend on maximizing their particular person and workforce’s success”
Not solely that, however AI seems set to develop help reps’ improvement alternatives, and open up completely new expertise and profession pathways. From bot orchestration, to data base administration and optimization, to product schooling – the much less time spent on repetitive FAQs, the extra time reps can spend on maximizing their particular person and workforce’s success.
AI is overhyped
We get it, we additionally had a wholesome dose of skepticism initially, and after the crypto bubble, who wouldn’t? However now that we’re seeing AI dramatically easing the workload of support teams, its energy and potential is not possible to disregard.
Fin can resolve as much as 50% of questions, immediately. With half your workforce’s help queries taken care of, consider the alternatives you might notice by way of optimizing your help workforce operations, enhancing your buyer expertise, and constructing relationships for extra seamless cross-team collaboration.
“AI is hyped. It’s hyped to shit. And but I believe the hype’s actual”
That’s why we imagine that AI in customer support is being hyped simply sufficient – not due to what it may well do, however due to what you can do with the time it provides you again. Our Chief Product Officer, Paul Adams, has labored on the heart of two large technological advances, and is satisfied AI is bigger than both.
“I labored at Google within the cell workforce once we launched Android… I labored at Fb on the peak of social a couple of years later. AI, to me, feels approach larger, approach larger [than those] and AI is hyped. It’s hyped to shit. And but I believe the hype’s actual.”
Clients aren’t prepared for AI chatbots
Clients have historically had a rocky relationship with chatbots, usually discovering them awkward, unintuitive, and unhelpful. However the will to self-serve is powerful, and contemplating the ability and potential of AI chatbots, 73% of support leaders imagine that prospects will count on AI-assisted customer support within the subsequent 5 years.
That signifies that the following couple of years mark a vital juncture for customer-chatbot relationships. Our recent user research confirmed that prospects are already far more assured that an AI chatbot will perceive their queries, and in the end have the ability to assist them.
When you’ve been enjoying round with ChatGPT, you’ll already expect AI chatbots like Fin to converse in a extra pure, human approach than we’ve been used to prior to now. It’s instantly simpler for a buyer to convey their level, set up an understanding with the bot, and get a solution they’ll perceive and query additional if wanted. Not solely that, if Fin doesn’t perceive what a buyer is asking, it may well dive deeper, asking clarifying questions to determine precisely the best way to assist.
Hallucinations make AI unsuitable for customer support
Generative AI’s trustworthiness was a significant query when ChatGPT first got here on the scene – and one of many greatest challenges we confronted when attempting to construct our personal AI chatbot. Regardless of the unimaginable potential of generative AI, there was no approach we may threat a chatbot confidently offering believable however incorrect info to our prospects.
“Now, firms throughout a variety of industries are discovering methods round generative AI’s greatest flaw”
Nevertheless, because the weeks went on, we began building the guardrails and security measures that kind the inspiration of our AI chatbot, Fin. To cite Fergal Reid, Intercom’s Senior Director of Machine Studying: “We’ve accomplished a number of work to make use of the big language mannequin to be conversational; to make use of it to know a assist heart article you may have, however constrain it to solely giving info that’s in an precise assist heart article that you just management and you can replace and you may change and you may edit.” And it’s not simply us – now, firms throughout a variety of industries are discovering methods round generative AI’s greatest flaw, and making their prospects’ lives simpler.
If Fin can’t reply a query based mostly on the help content material you present, it would admit that it doesn’t know, triage the shopper’s question, and seamlessly hand off to a human help rep. Consequently, you by no means have to fret about your buyer receiving a sub-par expertise and leaving the chat with rather less respect to your model. Both their question is resolved by Fin, or it’s handed to certainly one of your trusted reps for some extra TLC.