Conversational UX: The missing piece in your chatbot strategy

If given a choice between being on hold with customer service and having a query resolved by a few taps of a smartphone keyboard, most of us would pick the second option. It’s easy, quick and according to Gartner, 80% of companies will switch from native mobile applications to messaging by 2025.

Despite knowing this, many of the chatbots we encounter on a daily basis just don’t cut it. They lag, misunderstand simple questions and above all, don’t meet the standard of intuitive design that consumers expect.

This isn’t to say that the chatbot industry hasn’t evolved since its conception. We’ve come a long way from the clunky, barely functional chatbots of the late ’90s and early aughts. However, there’s room for improvement and innovation. The missing pillar? Conversational UX.

Help TechCrunch find the best software consultants for startups.

Provide a recommendation in this quick survey and we’ll share the results with everybody.

What is conversational UX?

Conversational UX has largely been ignored in the bot-building process. It’s an entirely new paradigm in this space, but it’s not a new hurdle altogether. Every new advancement in tech is accompanied by a discussion on how humans can interact with the the tech for better results. Technologists aren’t just tasked with making sure the products work, they must also devise ways to make the experience functional.

Though chatbots are largely meant to handle simple customer service tasks, there is an opportunity to scale both customer service and sales messaging.

Conversational UX is an emerging field, but despite the need for more intuitive chatbots, the industry as a whole is not spending the necessary time and effort into perfecting the experience.

Chatbots often function as glorified web forms, without any of the intuition or seamless integrations that consumers expect to see when interacting with “smart” platforms. But the industry has improved dramatically in recent years: Modern chatbot platforms have hundreds of pre-built integrations, which enable companies to connect their existing systems and tools to provide a secure, unified customer experience.

With every new interface, the goal is to make human-machine interactions better and result in a more intuitive experience for the user. Conversational UX presents a greater challenge because of the nuances involved with human language. It requires careful thought, empathy for the user and significant design considerations to carefully craft elegant experiences.

Let’s walk through a few ways we can make the experience better for the average end user.

Omnichannel capabilities

According to a recent study, the average American household has about 25 connected devices, including smart TVs, smartphones, tablets and laptops. This means we’re more connected than ever, with endless opportunities for companies to engage with their customers. But often, chatbots are fragmented across different touch points and can’t carry the historical data of past interactions with customers.

Consumers have come to expect and appreciate, continuous, intuitive conversations with the brands they engage with. Whether it’s reminders of items left in carts or questions about orders, ongoing communications across different devices result in more convenience and trust in the company. Omnichannel integrations ensure that companies only need to build one bot to deploy across channels.

Voice and tone

Although the conversation between the customer and the chatbot should be seamless and as human as possible, in order to maintain a level of trust, it should be clear from the jump that the chatbot is just that — a bot. Any hesitancy a customer experiences toward interacting with a bot versus a human employee will be mitigated if the platform is intuitive and straightforward.

The verbal voice or tone of the bot will vary depending on the company. If the product or service is business-facing, the bot will likely take on a more professional tone, while consumer brands have the freedom for a more lighthearted voice. The voice of the chatbot should align well with the company’s external-facing messaging and brand guidelines.

Multilingual options

When developing a chatbot, it’s crucial that the chatbot is programmed to understand and respond in multiple languages. Not only does this help companies reach global customers, it’s also valuable domestically — over 20% of U.S. residents speak a language other than English at home.

As the world becomes more connected, and products and offerings reach a wider audience, it’s critical that all potential customers feel comfortable and understood. A multilingual natural-language understanding (NLU) can be quickly deployed across geographies, and the bots can use self-learning to improve accuracy with every interaction.


A primary benefit of text-based communication is that the data is collected and stored regularly. Chatbots should be programmed to regularly assess feedback from customers and update in real time. Any hiccups in the communication process can be used as training data to improve efficiency.

For example, if there’s a pattern of customers selecting a specific option when communicating with the chatbot, the bot can automatically move the option upward to make it easier on the customer. But if customers rarely select a certain button, it can be moved lower or removed from the menu entirely. This data should then be automatically reported back to the organization to identify drop-offs and opportunities to improve.

Providing additional value

Though chatbots are largely meant to handle simple customer service tasks, there is an opportunity to scale both customer service and sales messaging. Chatbots should be programmed to provide proactive updates on delays, shipping timings and upcoming sales programs. Additionally, these bots can provide sales support by recommending other similar products.

Proactive support and upsells don’t just help increase sales and revenue, they also help build customer trust by providing value and anticipating their needs.

Communication loops

One frequent cause for frustration when it comes to chatbots is the seemingly endless loop of miscommunication that can occur when a platform doesn’t understand a customer’s message. When programming a bot, developers should ensure that the platform can understand and account for common misspellings, shorthands and slang terms. This is where natural language processing can be used to understand customer intent and queries, and respond efficiently with a high degree of accuracy.

Bots should also be programmed with sentiment analysis capabilities to ensure they pick up on text in all-caps, excessive punctuation or emojis — all of which may represent frustration or anger, at which point the conversation should be passed along to a human customer service associate.

Augmenting human workers

One common misconception about AI and chatbots is that they will take over jobs from human employees. But this couldn’t be further from the truth: Chatbots should be treated as a tool for a company’s human customer service associates, not as a replacement.

The chatbot can provide a customer service agent most of the information they need even before they have to speak to the customer. Once the customer is passed over to the agent to discuss the problem, the agent can immediately begin solving the problem without having to spend time gathering information. The data collected by the chatbot can also help pull deeper insights into the customer journey.

The chatbot market was valued at $17.17 billion in 2020 and is projected to reach $102.29 billion by 2026. Chatbots are not a trend; they are here to stay. If we want chatbots to earn their place in the market, we must ensure that bot developers are equipped with the right knowledge to improve the customer experience. Mapping all decisions back to the end user and their expectations is crucial and mutually beneficial to both parties.