Published on April 20, 2020
Building human-like conversations for chatbots
“Ok Google, transfer 10k to mom”
Well, yes!!! Now you can set up Google Assistant to let your customers make fund transfers easily.
You probably are aware of the various conversational assistants like Google Assistant, Alexa or Siri. You talk to them because of their ability to naturally understand your problems and provide you with solutions that you are looking for.
Ray Kurzweil, an inventor, futurist, and engineer at Google (who has a pretty good knack for making accurate predictions) predicts, that in about 9 years from now humans and chatbots will be indistinguishable?
We at Active.ai are crafting intelligent banking conversations. We strive to make our clients and their customers feel that through conversational AI we are not just providing them a platform to extend their banking services but we are getting them closer to their customers.
In this article, I am going to walk you through some of the delightful customer banking experience practices which we follow to empower and engage clients and their customers into meaningful conversations.
1. How should the chatbot talk?
Does your chatbot have a personality? If yes, then how is it a representation of your brand? The chatbot’s personality can be anything from being polite, friendly, upbeat or even funny!
Also, what capabilities should your chatbot have?
Our chatbots should be smart and quick in solving our customer problems like answering the FAQs, finding nearest branch, balance queries, fund transfers or bill payments etc.
Well… This is what most of our clients are looking for. The first thing your bot says defines how it wants to lead the conversation with the users.
Let’s see an example here…
Right: Chatbot has a persona and is offering the services to the customers.
With the conversation initiated on the right side, your customers will start to connect with the chatbot better.
Another important aspect of the chatbot is the verbiage it uses. Chatbot’s verbiage should be easy for the users to understand and MUST not contain a lot of technical jargon.
Use of technical jargon in the banking context
Excessive use of jargon is a BIG NO as it can be intimidating for the users. You definitely do not want to scare them!
2. Transactions via natural conversation
Why ask users to fill really long forms taking their banking details every time when you can use a natural conversation? This is exactly why we want to shift the mindset from a traditional application kind of experience where you have to navigate through a bunch of different screens, fill your data at a few places to complete a task or goal.
Let’s assume your customer wants to send monthly pocket money to her daughter. What should she be telling the chatbot?
Customer making a transaction via natural conversation
Now, let us break it down to see how are we classifying this intent.
- 10k: Amount to be transferred
- Tara: Receiver
- Pocket money: Remarks
3. Uncover hidden intents in between conversations
This is something which makes the conversational experience to be more humanized. How do you talk to your friend who is feeling low? “Hey! You sound very low, do you want to go out?”. You understand that your friend is lonely and nothing is better than a shopping therapy for her!
We do the same while crafting the chatbot’s responses. It should try to discover what the user may be looking for next based on their intention.
A human’s intuition and empathy are often a major factor when satisfying service requests or managing disputes with customers.
Suppose your customer has lost her wallet on her travel to Cochin
Your customer has panicked! What can they do!?
So here is how your chatbot can help to uncover a hidden customer intent.
4. Remind tactfully for better engagement
A nudge to the user asking them to pay monthly bills could help start an engaging conversation and make them save on late payment charges!
Also, you may want to keep them posted on the offers that they might get have based on their recent transactions.
Customers would get a reason to come back for another transaction.
Here is how we make this simple with micro conversations.
5. Fail gracefully
There is no such thing as a perfect chatbot. There can be possible user utterances where your chatbot cannot understand what the user has typed or is not trained to handle those utterances. In such a scenario let’s see how can you keep the conversation going.
What is it that the chatbot has failed to understand? If you would like to take in those utterances to extend the capabilities of your chatbot’s understanding, take in all that data, analyze them and train your chatbot to handle it gracefully the next time. Your chatbot is getting smarter day by day!
5. Take in feedback while they exit
After you have provided the banking service to your customers, it is time for you to ask them about your service. How was their banking experience? Did they face a problem which they want to report?
Remember! User feedback not only gives you the scenarios where your chatbot has failed but also an opportunity where your chatbot can be improved and made more efficient. Make an iterative approach to analyze and incorporate their feedback into the upcoming releases.
It could help you uncover real customer problems that you may have overlooked.
Not just that, you may be able to analyze what queries do your customers actually have. Are they using the chatbot to perform transactions? Ask some FAQs? or maybe just having a little conversation to see how the chatbot responds to it? We all like to play with the chatbot in crazy ways!
6. Speaking the language of the future
The future is always unpredictable but continuity in your services would help you bring a connection with your customers. In trying times, like the Global pandemic of COVID-19, when customers may feel lost, your on-going services can help to gain their confidence and reassurance that you are always there for them.
Conversational AI across different channels like voice, messenger, etc.. is one such technology that can aid in making the banking business more robust. It ensures continuity of your services so that the customer does not feel disconnected.
So remember, even if your customers are sitting back and working from home, they can just send a text message and complete their banking transactions!