Introduction

Hi everyone, Mark here from Voiceflow, and I'm thrilled to share another template that will surely help you get you started with your next chatbot experience.

Introducing the Banking Assistant Chatbot Template - a fully customizable banking flow that enables customers to complete two main functions: transfer money and check their spending history. It can also complete a handful of sub-functions or FAQs such as finding the nearest branch, checking your account balance or providing general help.

How can I use this template?

Although chatbots are evolving fast, they can still leave customers feeling frustrated or confused. To ensure customers enjoy a seamless experience, businesses like banks need to continue improving upon their multi-channel offering and the ability to tailor personalized and convenient experiences across them.

By utilizing the Banking Assistant Chatbot as a starting point, you can build out a variety of great customer support chatbot experiences, such as a Hotel concierge, Credit Card assistant, or a product tracking & information chatbot for your online retail store. All of which can maximize customer service, minimize customer service costs and increase self-service channels.

Topics we will be covering

In this video, we'll cover a few main points while walking through the template design. This includes:

1. Using top-level intents to create flexible menu options
2. How to set up commands to support FAQs
3. How to use the Custom Action step to simulate an event
4. Creating utterances and required entities to fulfill the user's intent
5. Testing the experience

Design Walkthrough

So the way we've designed this template is firstly using open intents - or a collection of phrases that capture a user's intent and then trigger an action. ****And so, when the project hits this prompt step at the start of our experience, our chatbot is going to stop and listen for the user to say something.

Now the first question we ask, is, "what can I help you with today?" A user can say transfer money or check their spending history. Once they actually say this, it'll immediately match it to the transfer money or spending history intent, which puts the user into a closed system where they are essentially pulled down a deterministic path. And so this is the basic structure of our design.

To understand the components and flow of this experience, let's head back to the start block.

As you can see, we've used two commands within our experience: General Help & Account Balance. Each command is linked to an intent and sub-flow which can be invoked anywhere in an experience. For example, if a user says account balance they will immediately activate the account balance conversation flow. Since the account balance intent is housed within a command, once the user finishes the flow, they will return to where they were previously in your project.

Within the flow of the Account Balance command, we've employed a set block to set the value or current balance in the user's account to a fixed number. This was done to quickly simulate what this experience might be like, however, to accurately retrieve this data from a user's bank account, you will need some sort of API integration between the user's bank and your experience.

Now after the set step, we've completed this flow by incorporating custom visuals, responding with the account balance using the balance variable (which we set in set step above), and then ended with a prompt step. We've also employed a confirmation intent to help redirect a user back to the home flow and a sub-command in case they ask where the nearest branch is.

Let's now head back to the Home Flow.

Now, after the start block, we hit a Custom Action Step. These steps are used to simulate an action or event in your experience. This could be opening a car door or fetching credit card info — actions a developer could implement during the design hand-off. Here, we're using the Custom Action step as a placeholder to simulate an API request that pulls real-time data from the user. APIs provide a powerful means to pull important information from an existing CMS, database, or internal tool so you don't have to worry about storing customer information in your experience.

When you open up the editor, you've got a few options, like setting up multiple paths, adding labels, or even creating a default path for your conversation.

After the assistant asks, What can I help you with today? There are two main intents the user can trigger: Transfer Money and Spending History.

To satisfy the different ways a user might ask to transfer money, we've included a variety of utterances along with 3 key entities: contact, amount, and delivery date. Most importantly, we have made sure that each entity is required - meaning this info must be provided in order to complete the step. That way, if the user fails to fill any entities during their response, we can continue to ask the user (using a response utterance) for that information until it's provided.

From there, the user can send, edit or cancel their transfer using the button step. On send, the assistant then confirms the transfer using the variables filled by the user. We've also used custom visuals to accompany the confirmed request and a prompt at the end of the conversation in case the user wants to perform another task.

Finally, if the user triggers the spending history intent, they will head down this branch here, where — using this IF statement — they will be filtered down a specific path based on whether they provide a spending category in their response. You can consider an IF step like a crossroad in a conversation. From there, we've employed button steps on either side of these branching conversations to further qualify the user's intention. The right information is eventually spoken back to the user using variables which eventually leads to an assistant response and prompt step where the user can decide how they'd like to proceed.

Once again, it's important to remember that we're using this template to simulate a banking chatbot experience, and so we've employed the use of fixed numbers as you can see here, and here. When building your own custom chatbot experience, be sure to connect your own custom APIs in order to simulate a 1:1 experience where you can collect real banking data from your customers.

Let's now see this chatbot experience in action by clicking on share prototype and then hitting open link in new tab.

Prototyping Experience

Chatbot: Hi there, and welcome to Mobility Bank
Chatbot: Let me know how I can assist you today
User: Whats my balance
Assistant: Your account balance is $6420
User: Transfer Money
Assistant: Who should I send this to?
User: Mike
Assistant: How much will you be transferring? 
User: 50 dollars
Assistant: When should I send this? Please enter a specific date, for example: January 1st.
User: March 3rd
Assistant: Ok, sending $50 dollars to Mike on March 3rd. Is this information correct?
User: Yes
Assistant: {Visual} We've scheduled $50 dollars to be sent to Mike on March 3rd.
Assistant: Is there anything else I can help you with?
User: Check my spending history
Assistant: How many days of transactions should I show? 
User: 7 days
Assistant: You have spent $1301.82 in the past 7 days.
Assistant: Would you like more detail? 
User: Yes
Assistant: {Visual} You have made 12 transactions in the last 7 days
Assistant: Anything else I can help you with?
User: No
Assistant: Thank you for banking with us today. We hope to see you soon! 

Conclusion

From ride sharing to retail to SaaS products, businesses of all sizes are developing chatbot assistants to deliver seamless and personalized experiences users can count on. We hope this template serves as a strong starting point for your next chatbot experience. If you want to learn more about conversation design or how you can get started with Voiceflow, check out our templates library. It serves as a great resource to inspire new ideas and is entirely free.