The idea that a computer can ‘talk’ to a human has been around for decades and is one of the most common sci-fi tropes. Today, we’re seeing fiction turn into reality with AI-powered chatbot and voice assistants slowly changing the way we live, work and shop. More and more businesses are starting to explore how a conversational user interface can impact customer experiences. From commerce and support to process automation and enablement, new possibilities are emerging every day. However, before we go into the details, first things first.
What is a conversational user interface?
A conversational user interface (CUI) refers to an interface design which enables people to interact with a computer on ‘human’ terms. The UI mimics human modes of conversation via speech or text, making it easier for users to interact with a system and eliminating the learning curve inherent to traditional graphic user interfaces (GUIs).
There are two major types of conversational UI in use today - voice-enabled bots and text-based bots. Examples of successful CUIs include voice bots like Apple’s Siri and Microsoft’s Cortana and chatbots like Mitsuku and Cleverbot. The third kind of CUI blends both voice, text and graphical elements to offset the drawbacks of each and create a more focused, immersive user experience.
Creating a system that can accurately simulate human conversation while delivering value to the end-user is no mean task. Conversational designers have to understand and apply concepts from visual design, UX writing, natural language processing (NLP), interaction design and multiple other disciplines to create the optimal user experience. By outlining the logical flow of conversation, mapping potential input with responses, and defining the character persona, conversational designers seek to create a seamless Human-AI interaction.
Rule-based Vs AI-powered CUIs
It’s important to understand that the mechanism behind conversational UIs falls into one of two categories - rule-based and AI-based.
Rule-based CUIs are relatively straightforward to design and involve mapping user responses to a specific output. If the user submits a response that does not align with a designated system response, the UI displays an error message or redirects the user to a human executive.
AI-based CUIs are more complicated affairs. These CUIs use a combination of natural language processing and machine learning to try and understand the context and purpose of the user input, before generating a curated response.
Conversational UX vs. Conversational UI - Is there a meaningful difference?
A lot of people tend to use the terms ‘conversational UX’ and ‘conversational UI’ interchangeably, but the difference between the two is both nuanced and enormously important.
Conversational UX (CUX) refers to the entire experience between a user and an organization being defined by a conversation. Conversational UX designers are involved with market research, defining dialog flows, understanding the pain points of the customer and mapping out a relevant user experience from a conceptual standpoint.
The Conversational UI designer then takes over. UI design combines elements from both visual and interaction design to bring the UX to life via a specific interface. Will the user have buttons to advance the conversation, or will it be done via more complex means of input? If the buttons are clicked, will they generate an audio cue or a visual cue or both? What kind of transitional animations will move the user from one page to another? These are all questions that are tackled by a CUI designer.
Both UX and UI design, however, can be limited by the features of their Artificial Intelligence platform - the foundation on which any intelligent conversational interaction system is built.
How to build an omnichannel conversational experience?
When designing a conversational UI, consider what problems it is intended to solve. Are you looking to improve sales on an eCommerce platform, build leads for a landing page or facilitate better organizational efficiency?
Secondly, take stock of your user needs using a combination of quantitative and qualitative data collection methods including surveys and interviews, and by analyzing your database of customer data. This exercise will help you create a CUI that is contextually relevant to your customers’ needs.
One conversation - many channels
The next step is determining how your CUI will interact with customers. For example, a text-based CUI for eCommerce could ask questions that require the user to input text, choose from a selection of responses or a combination of the two. But if you were designing a voice-based interface, you might prefer a different kind of interaction architecture - one that offers both, fewer menus and a high degree of natural language understanding.
Platform and channel integration
Creating an omnichannel CUX will require a platform that allows seamless integration with multiple channels. A good example of such a platform is Senseforth’s A.ware, which integrates with major consumer messaging applications like Twitter, Facebook Messenger, WhatsApp, as well as enterprise channels like Slack and Skype. A.ware also uses six different engines to drive advanced natural language processing, contextual analysis, conversational intelligence, machine learning, actionization, and analytics, making it an end-to-end chatbot design platform for both consumer and enterprise needs.
Customer touchpoint mapping
Once you’ve settled on a platform, it’s time to map out the omnichannel customer journey. For example, a retail customer might start researching products on your website and then switch to a shopping app, while post-sales conversations might happen via a social media platform or SMS. From here, it becomes necessary to identify which channels can be automated or taken over by a CUI and how they will deliver value to each customer.
Mastering dialog flow
Architecting the flow of dialog is also key to creating a relevant conversational experience. Start by jotting down all the queries that could be triggered during each phase of the customer journey. Then you can begin tailoring dialogs to each query and creating diagrams of different conversational flows. Note, an individual usually speaks, types and reads differently, using different language styles for different communication formats. Craft responses that align with either voice or text (depending on your CUI) to create a more human-like CUI persona.
Teaching AI about people
Even with advanced machine learning and natural language processing suites, most CUIs need a database of baseline customer queries to learn from. Not only does this database help your CUI learn about and respond accurately to the spectrum of customer queries, but it also allows you to create a persona that aligns with user expectations.
Back-end integration & interchannel communication
Finally, consider how your channel-specific CUIs will communicate with your back-end systems and each other to create a seamless omnichannel experience. Currently, API calls and platform-as-a-service packages are considered cost-effective and efficient tools to integrate multiple channels, including mobile applications, webchat, and social media platforms.
How conversational experiences can drive business?
Most consumers still view conversational interfaces as customer support or personal productivity tools.
And this perspective is not entirely misinformed. By and large, the potential for CUIs to boost revenue, nurture leads and assist humans in complex tasks has not been fully explored.
However, design and technology advances are quickly helping CUIs surpass this role, and move into enterprise communication, eCommerce, personalized marketing, content curation and delivery, and lead generation.
The rise of conversational commerce
Created by Chris Messina of Uber, the term “conversational commerce” refers to a concept that enhances the online shopping experience via chatbots or voice bots. Conversational commerce aims to move the entire eCommerce experience on to a conversational interface, where the user interacts with an AI or rule-based shopping assistant, much like they would speak to a store associate at a brick-and-mortar retail outlet. Conversational commerce allows users to access customer support, locate product information, create wish lists, read recommendations and reviews, and ultimately purchase products via a messaging or voice interface.
According to Facebook, 61% of US consumers have messaged a brand in the past 6 months, while the Narvar Consumer Report 2018 states that 29% of US online shoppers plan to use a chatbot to make purchases. These revelations are indicative of a growing trend toward automation and AI deployment in consumer-facing applications across developed, tech-empowered economies.
A conversation about sales
Today, conversational commerce largely uses AI and analytics to create a personalized shopping experience for each user.
Users today are using messaging apps more often than social networks - according to multiple industry reports from TechCrunch and Facebook, messaging app usage has surpassed social network usage by almost 20% over the past 5 years. Given these developments, businesses now have the opportunity to create a real-time, bi-directional line of contact with their customers via conversational interfaces.
According to a study by Usabilla, 65% of users would prefer talking to a chatbot rather than waiting in queue for a human support executive.
For eCommerce operations, great customer support is critical to growing the user base, cross-selling and upselling, and growing brand equity. A simple rule-based chatbot can unburden your customer support teams by tackling simple customer queries, while humans can address more complex customer problems. If you don’t have a dedicated, round-the-clock customer support operation, you can use bots to log tickets and respond to users during off-hours.
Chat and voice bots are also a great way to drive new lead generation initiatives. Consider replacing your landing page or website contact form with a chatbot that engages users and collects their contact information, along with purchase queries and intent markers. In essence, the chatbot offers an automated mechanism to qualify and nurture inbound leads, saving you time, resources and most importantly, driving up your sales numbers.
One of the most famous examples of a business driving customer engagement via a CUI is SBI Cards. After deploying its chatbot ILA via a personal banking app, the bank registered a multifold increase in the number of daily users on the app, thereby helping the company drive top-line revenue growth.
Another great example that takes a more indirect approach, is Insomnobot-3000 - a CUI designed to hold late-night conversations with people who have trouble falling asleep. Introduced by Casper, a firm that specializes in sleep products, Insomnobot-3000 is able to deliver a valuable and memorable service to consumers while driving lead generation and gathering a ton of relevant psychographic information about Casper’s core customer base.
Is there an enterprise function for CUIs?
In the enterprise context, conversational interfaces have the potential to improve workplace efficiency via simplification and automation.
Instead of having to maneuver through a new graphical user interface for different applications, a conversational interface allows users to simply voice or type out the desired function. Similarly, CUIs can effectively take over multi-step tasks that previously required a human assistant. For example, delivering a command like “Send the list of approved project candidates to Paul at 5 pm” - this task would likely take a human assistant longer to execute and would consume more man-hours when compared to an AI-powered CUI.
And of course, while the AI would need the right programming to understand the context of your statement, they can be deployed immediately without the training and adjustment period that a human resource would require. This ability to interpret natural language input and execute complex tasks on that basis has long been a key limiter for AI - one which has thankfully been overcome by NLP algorithms.
Multi-tasking also becomes easier within the ‘conversational enterprise’ as personnel can deliver commands and instructions to multiple applications without having to manually click buttons and navigate through different GUIs. If we extrapolate and scale this concept, we can see how enterprise staff could soon orchestrate commands and actions through their entire application landscape in multiple geographies, simply by having a dialogue with the CUI.
By combining all the above capabilities, enterprises can create a low friction work environment where applications and software functions are open to all users, and not just IT specialists. Additionally, the use of CUIs to execute enterprise functions could help increase internal compliance, drive informational transparency and improve overall user satisfaction.
Will RPA integrate with Conversational AI?
The rise of CUIs as viable business tools also begs the question - how will CUI integrate with robotic process automation (RPA) to create meaningful business impact?
A great example can be found in the insurance industry. Usually, the user journey to initiate and receive insurance claims involves several business systems and human approvals, where an error in one system can cause severe process bottlenecks.
Consider a situation where a customer needs to file a claim for broken consumer electronics. By integrating CUIs into the claims management system, the insurance company can accelerate this whole process. The customer can simply upload documentation, video or images. In turn, CUI can validate the claim faster. The CUI can also deliver pertinent information to the customer such as the policy details and time to claims resolution and can even enhance the customer experience by aggregating and displaying discounts on different replacement hardware models.
At this point, RPA could greatly streamline the rest of the customer journey by automating the validation checks, configuring policy renewal, and even settling the claim via direct reimbursement. By applying this paradigm to other industries, we can easily see how the intersection of CUI and RPA could completely change how we apply for loans, book holidays, find vendors and initiate business contracts.
What’s next for CUIs?
Businesses across the world are seeing an increase in the adoption of CUIs over other communication formats like email. But there are still significant strides to be made.
A study by Userlike shows that 59% of users would prefer it if chatbots were more human in their interactions. This can range from injecting humor into conversations to responding better to conversational cues. In essence, users are looking for a conversational interface that offers more personality and verve than a standard set of robotic responses. Chatbots also suffer from a lack of dynamism and language processing failure, especially in the area of context analysis - however, this issue is slowing being overcome as natural language understanding technology continues to develop.
As CUI technology advances, we are likely to see brands adopt and configure CUIs as branded digital assistants that handle everything from sales and marketing to customer support and lead nurturing. Similarly, the managed IT services sector is just a couple of years from implementing powerful AI assistants that can efficiently handle workloads from both internal and user-facing divisions.
In fact, research and consulting firm, Ovum, expects that there will be over 7.5 billion digital assistants by 2021. That’s more than the global population, and definitely a positive trend indicator for organizations looking to deploy CUIs.