How to design chatbot conversations to wow customers!

How to design chatbot conversations to wow customers!

SnatchBot Team SnatchBot Team, 10/03/2020

How to design chatbot conversations to wow customers!

When it comes to building a chatbot there are many ways and inputs on how a business can think about and design the conversation, but taking a customer-centric approach is key for successful chatbot conversations, along with a strong design.

The hype and interest of chatbots can see many teams and designers carried away with the excitement of the task, overlooking the key role of the customer or audience at the heart of any conversation, with many different types of chat available, and a wide range of sentiment and expression that can add clarity or confuse the user.

From marketing to customer support chatbots, public information bots to internal HR or office services, designing a bot that delivers for the business and the customer with strong conversations is key, all using smart chatbot conversations to deliver a practical and useful service.

In a few short years, chatbots have grown from simple conversations to increasingly smart tools to help link businesses and customers. While the underlying cloud technology might sound deeply scientific or in the realm of big data boffins, the tools to build good conversations are remarkably easy to master.

Designing a Conversational Chatbot

Transposing the traditional conversations we are all used to having in person, on the phone, via email or chat to a chatbot sounds easy, but requires some thought to get the message and interactions right. Bot tools are plentiful from the giants like Watson and Google to bespoke vertical market tools, but whatever the back-end service, every business needs to plan and design when it comes to building valuable conversations for a company.

Dealing with customers, clients, patients, partners and other people may come as second nature to sales, receptionists and customer support agents, but leaving a chatbot to cope with the many nuances and possibilities is quite a stretch.

When it comes to building the bot, you need a platform to create it with, destination platforms to host it, like Facebook Messenger, or within your app or website. But beyond that, you need to ensure the conversation is appropriate, is well designed - using a flow map - that meets the needs of those on both sides of the bot and delivers everything that is needed for a successful outcome.

Designing the conversations requires and understanding of how the chatbot requires structured data, while users need helpful conversational cues, are encouraged to give pertinent information. In between the bot can use persuasive language to encourage positive outcomes, and identify when a user is struggling or stressed and offer options to calm them down and create a successful resolution.

This level of fidelity is a long way from the early bots using “Do you want A,B,C?” and following a strict tree of conversation options. You can still design bots like that for simple use cases, but

people often need empathy and bots need to appear helpful and natural (like Siri or Alexa), rather than bland and direct automatons.

To start with, your bot can have a persona, just like Alexa, to help customers get over the artificial nature of a chatbot. A simple “Hello, I am Boris, how can I help you?” helps break the ice and sets the conversation as a typical one. Boris, or whatever you call your bot can represent the company brand or tone of voice, or can have added character depending on the nature of the conversation to help engage.

With your character as a starting point, you can then flesh out the three key areas of the conversation:

  • The introduction “how can I help you?” and perhaps identifying the user. 
  • The main part of the bot - solving their problem or delivering information.
  • And the exit point, ensuring a successful outcome and getting feedback or completing a sale.

Within these parts there are various elements that need resolving to reach a successful conclusion:

  • Intents - what we want to achieve, make an appointment, buy, sell, reorder or cancel
  • Entities - people, things, accounts, products, services, appointments
  • Motivations - Make a change, get something cheaper, seek help.
  • Queries - Can I do this? Can you help with that? 
  • Break points - I want to talk to someone, that’s not helping
  • Confirmations - Is this correct?

These can be built into a structured bot, although that structure can get very long and convoluted depending on the conversation. Or, it can be developed through AI training and recognition, using and outline and previous conversations to build a chatbot that reacts in the appropriate way to common inputs and learns how to handle new queries.

Choosing suitable manners with which the bot operates will it provide the appropriate “ tone of voice”, while focusing on the quantity and quality of conversation, and ensuring the context of the chat will help build something that flows naturally for the user.

Between the conversational script, the smarts of a conversational artificial intelligence, and the bot creator designing natural and useful interactions, any chatbot can be engaging and help meet the needs of the audience, no matter what is thrown at it, creating a positive customer experience that will satisfy them and see them use the bot more than other methods of communication.

Be aware that while confirmation is essential “is this your contact number?” people will get very bored with bots that constantly ask for confirmation, so build this into the bot at salient intervals rather than falling into a repetitive trap.

As bots become increasingly complex, they can be tested in an artificial intelligence conversation simulator to highlight flaws or potential issues that customers would have, but a good dose of personal testing by people outside the business is also essential.

Building a Conversational Growth Strategy

For marketers, the rise of the conversation in business gives that chatbot greater importance than in other areas. While most bots can save time and some money for the business, a marketing chatbot can drive the sales conversation, help engage with awareness and encourage upselling and sharing, all highly valuable to the marketing organisation.

Instead of throwing adverts at people in the hope of gaining their scant attention, encouraging a relationship by designing conversations helps to build a relationship that lasts across the lifetime of the business. Chatbots are perfect for this approach, saving time but providing insights and automated engagement across multiple points, including texts, apps.

By providing the right information to the prospect or customer at the right time, it helps guide them towards a purchase, can encourage aftercare and other inquiries. And chatbots are there 24/7 to deal with these queries. With the right personality, they can help build the brand, become part of smarter advertising and provide live feedback to the business, with a useful way to better understand your customer.

Chatbot Design Technology and Best Practices

Designing a conversational experience might sound new and scary, but there are already plenty of best practices to follow, while the underlying AI technology like natural language processing and understanding all help to make it something that any business can achieve.

Fortunately, as chatbots are a fast-maturing technology, there are also plenty of chatbot sample conversation examples that can be used as a starting point for your business or service, and many solid examples of good bots in public use across most markets. 

The use of conversational AI and user interfaces are key to building a smart and successful chatbot. Conversational AI applications help move the conversation from one repetitive chat to help build long-running interactions that can handle multiple needs and grow with the business. So, the first chat could appear on their website, once a sale is made, it can check in with a satisfaction query using emojis if appropriate, and after an appropriate time can provide deals on upgrades or consumables. 

The conversational AI can also help navigate unexpected requests or learn from changing user interactions. They can engage customers at the appropriate contact point using SMS text, voice, emojis or another interface, whatever is most suitable for that time. They can also help avoid dead-end questions that can kill a conversation or lose a customer, and provide key feedback or metrics to the designers on how to improve the bot as users highlight what they want the bot to do next.

Finally, ensure that privacy and security notifications are available, and that your bots meet the required standards to protect user identity and data. And, there should always be an easy way to find an alternate contact point, be it an email address or live chat feature for customers who are struggling to cope with the bot.

What to Avoid When Designing a Conversational Chatbot

Much technology design is focused around meeting the business need, and it can be easy to forget about the customer. Add plenty of customer experience testing to ensure it meets their need and there are clear routes for them to provide feedback.

While there are some cases where funny chatbot conversations are a valid use case, seek to avoid humour for its own sake, or building bots that can stray too far from the main conversation.

Business people love to talk about their products or service, and that can be reflected in bots with long sentences, convoluted decision trees and too much information provided. If a bot conversation feels too long, provide links to the broader information for people to view in their own time.

Finally, remember that chatbots and AI are only tools, they need to be tested and checked regularly. Never treat a bot like a done-and-dusted project, as that will quickly put off users. Alongside that, ensure it is regularly promoted and updated with new elements or responses. 

Winning With Your Chatbot

Whatever your business, the keys to creating a good bot are to start off with simple chatbot conversations that can expand to meet growing business and customer needs. Then add the smart AI features that help expand the bot to act an increasingly valuable business tool that will play a key role in your future digital business.

Keep an eye on upcoming AI bot and chat technologies, and new ways they are being used to improve consumer or user interactions. And keep updating the bot on a daily, weekly, or monthly basis to provide accurate information and to ensure that it isn’t falling out of favour with customers.

From startups to industrial giants, growing numbers of businesses, governments, charities and other organisations are using bots. You can monitor their progress to see how the landscape is changing and how your company can better use bots. And as consumers get more used to seeing bots, they will expect to be able to use them, especially out of business hours, or in their own language as the digital world shrinks and bots become the de facto voice of most businesses.