AI for Customer Care: Addressing the 3 Key Industry Challenges

Gone are the days when artificial intelligence (AI) was a mere buzzword. Today, AI is transforming our lifestyle. AI enables machines to perform a myriad of tasks with minimal human interference.

Whether in sales, marketing, or customer service, AI has revamped business processes. According to a recent Gartner report1, 55% of established companies have started making investments in AI or have prioritized it for their immediate plan of action.

The applications of AI in customer care include:

  1. Assisting customers
  2. Providing basic information about products or services
  3. Identifying issues faced by customers
  4. Processing information and learning from it
  5. Determining customer behavioral patterns
  6. Sending notifications with important updates
  7. Mitigating cart abandonment and complaints
  8. Rendering real-time support in the form of FAQs, reports, and help forums
  9. Keeping track of user preferences
  10. Sending suitable solutions and personalized offers

Chatbots, also a part of AI, have a similar use case.

Take a look at this pizza ordering bot, a part of the customer service.

Artificial Intelligence

Despite the vast set of applications and the ability to offer uninterrupted service, AI can be a double-edged sword.

In 2018, Tencent, a Chinese multinational investment conglomerate fund launched “Zhiling”, an AI-powered translation engine. The bot performed live transcription and interpretation at its launch.

“Zhiling” was to demonstrate the ability to transcribe and interpret speeches in real-time. But it bore the brunt of social media jokes and backlash when the engine went haywire and didn’t serve its purpose. Certain words were generated needlessly and repeatedly. The bot got confused when the speakers spoke unstructured sentences. The result was:

While AI continues to grow and evolve in its capabilities, there are certain key challenges of confusing the user instead of helping using AI in customer service, that are yet to be addressed. In this article, I present potential solutions that can help overcome 3 major shortcomings.

#1 Lack of emotion and empathy

Customer queries are instantly resolved as a result of automated customer support channels such as live chat, social media, instant messaging (chatbots) and automated phone calls. As consumers, we don’t always require human interaction, we just need an answer to our query.

From offering greetings, bots can solve issues such as answering FAQs, conducting surveys, collecting contact details, sending out promotional offers, and explaining general rules. However, the main risk of AI is that there is an inherent inability to gauge and understand the tone and context of a conversation.

Bots lack the element which is the core function of human interactions – emotion. This is because bots today aren’t as intelligent as humans and they may never be. They’re built on decision tree logic. They respond to specific keywords that they identify from the user’s inputs. AI functions on the principles of science and algorithms, making it difficult to keep the human element intact. There is no human reasoning in AI and as a result, it comes off as an insensitive system that has no emotion or empathy, which are the driving forces of customer service.

Take a look at this conversation. The user is looking for assistance to book tickets for a trip to London, but the bot ends up confusing the user instead of helping.

Only a handful of bots have linguistic and natural learning capabilities.

Solution – Let customers know they’re talking to a chatbot and give them the option of talking to an agent instead.  A human agent should take over when the chatbot misses the conversational context.

For example, if the user wants to know their bill amount, the chatbot can answer. If they want to know their last billing date, the chatbot can answer. But when they ask, “Why was I overcharged?”, a customer support agent needs to take over.

#2 Inability to Localize a Conversation

AI is capable of increasing the productivity of support agents and ensuring that queries are dealt with as efficiently as possible.

Offering customer service in the major languages spoken by your customers is called translation, and ensuring all cultural nuances are taken care of is known as localization. For example, if you are selling a product in the US and the UK, you would have to change the dialect, the currency, the representation of the time and date, and the payment gateways.

For instance, Spanish is spoken by 400 million people in 20 different countries. The dialect of Spanish in Latin American countries varies vastly from that spoken in Spain.

So if you want AI to cater to customers in the foreign language, it is not as simple as translating the text. Aside from changing the script, the website will also have to be culturally compatible with the target audience.

Solution – This challenge presents a great opportunity for brands wanting to go global.

Improving localization is a great way to connect with customers in a simpler way.

Evernote launched in China under the name  “Yinxiang Biji2”. The website layout, name, payment gateways, visuals, and customer service whereas localized and made available in Mandarin.

Solution – Companies have to create a contextual experience for their customers. Speak to them in their language, answer the right questions, take their culture into consideration,  and they’ll become your customers for life.

You could hire a localization service provider to implement multilingual customer support. This way, you would not have to coordinate with multiple translators or language specialists. The end product that you get is content that is translated by native speakers taking into consideration the context and culture.

#3 The Cost of Development and Maintenance

 The cost of building a chatbot depends on its functionalities and your business requirements. Let’s talk about the factors and variables that influence the cost of chatbot development.

The industry vertical: If you are adding AI or a bot to healthcare and financial portals, you will have to integrate security protocols. This will add to the cost.

B2C platforms such as beauty, e-commerce, and travel require a higher amount of customer engagement. More dialogues will be required, which will increase the overall cost.

The interaction required: The cost of developing a chatbot is low for a menu-delivered approach. The user can select from predefined options.

If you are including natural language processing (NLP) to recognize the tone and emotions (for voice-based AI), the complexity and the time taken to build the bot will lead to an increased cost.

Capabilities required: Sending updates and FAQs require a lower cost of development as compared to training AI to book tickets or take orders. The latter requires deeper integration with the backend code.

Solution – The two options available in the market are:

Build your own bot: If you have developers in your team or the budget to introduce programming capabilities, you can build a bot according to your needs and preferences.

The downside here is the high cost of development required to build this. Ongoing maintenance becomes your responsibility.

This could cost you $25,000 to $50,000.

Use a chatbot builder: Pay a monthly subscription for a platform which has a prebuilt chatbot functionality or lets you build one.

For a monthly subscription fee, they give you a framework to build upon and support and maintain the bot.

These platforms start their pricing from $3,000 to $5,000.

The cost of a chatbot depends on your industry, company size and the requirements the bots fulfill.

Concluding Thoughts

Numerous brands have added AI in their customer support arsenal. The tricky part here is to decipher the queries that can be solved by AI versus those that require human support.

If you’re adding a chatbot to engage with customers, take into consideration the following points:

  1. Be careful that your customer service does not lose the human touch.
  2. Create canned responses for commonly asked questions as the first line of support.
  3. Allow helpdesk agents to focus on solving critical issues that require human intelligence, empathy, and emotions.

We would love to hear about some of the challenges you’ve faced while implementing artificial intelligence for customer care. Do drop us a note in the comments below!

1 –
2 –!

Leave a Reply

Your email address will not be published. Required fields are marked *