Across industries, companies are discovering the potential of conversational bots to help automate and streamline activities, improve enterprise productivity, and boost employee and customer engagement. While the earliest versions of conversational bots were simple response platforms, today’s AI-powered bots are much more powerful and will only become more sophisticated and capable in the coming years.


What is conversational AI?

Conversational AI is a form of Artificial Intelligence that allows people to communicate with applications, websites and devices in everyday, humanlike natural language via voice, text, touch or gesture input to enable a two-way digital customer experience. The first one is that the chat software or the chatbot must be lightweight and AI-powered. It is not a standalone app but is usually bolted on to another commonly used chat or messaging application. It could also be a lightweight skill or a capability on a commonly used voice-activated device. The software is AI-powered in that it not only responds to commands but is also able to infer, learn and make decisions like a human would. Without this inference capability, the bot will not be able to carry an intelligent conversation and will be no better than the legacy Voice Response Unit (VRU) that quickly transfers the conversation to a human. The second concept is that the bot should support both text- and voice-based interfaces as customers are familiar with both now and like the ability to seamlessly transfer between the two interfaces at their location and surroundings. The third concept is that it should promote an intelligent two-way conversation, otherwise the customer will be frustrated and either will request a transfer to a live agent, or worse yet, will go elsewhere to satisfy their needs. Hence, conversational AI, if done correctly, can improve the customer experience due to intelligent two-way interactions; provide unique data insights to the organization about the customer as the conversation helps capture nuanced customer data, like their needs and desires; and help improve operational efficiency by finishing more of the interactions without needing human intervention.

So, what are the methods of conversational interactions  that are available today? We have seen four broad categories of conversational assistants.

Messenger Apps– These are virtual assistants available on messaging platforms. The most common  and the widely used app is Facebook Messenger or WhatsApp. Facebook Messenger has a robust bot development framework that  is used by several organizations. Another common app  is Slack that is used by teams for collaboration and has  a robust bot framework as well. WeChat and Telegram  are apps that are more common in Asia, Europe and the Middle East and can be used to create chatbots as well.

Voice Assistants – These are general purpose assistants that are activated by voice and are available on mobile  and voice-activated devices. Alexa by Amazon is the most popular voice-activated device, though Google Assistant  by Google is also popular and gaining market share. Siri  by Apple and Cortana by Microsoft are voice assistants used on mobile devices.

Mobile App Assistants These are usually  text- or voice-based and deployed as an add-on or an extension to the organisations mobile app.

So, using one or more methods will improve the customer experience as they will enable intelligent two-way conversations on platforms that the customers are accustomed to, provide unique data insights from the conversation to the organizations and help improve operational efficiency by answering and solving more of the customers’ desires through the conversational channel.

So how can a financial institution build a conversational AI strategy? There are three broad steps that an organisation must undertake  to build a conversation AI strategy:

Understand where your customers are – It is important for organizations to understand what their customer touchpoints are and how customers move from one touchpoint to another. Organizations do this by plotting out user journey maps. These maps help outline how the organization's customer interacts. It is a good way of exposing friction points in the user journey and provides  an opportunity for the bank to eliminate or mitigate these  as it works to improve these user journeys and make them suitable for conversational AI.

Determine what capabilities to offer Once the organization has plotted out the user journeys, it is important to strategize and prioritize what capabilities should be offered to its users through a chatbot. These could be existing capabilities or even net new capabilities that are conducive for conversational AI. It is important that the organization employs a crawl, walk, run approach and introduces capabilities in a phased manner with lessons learned in each phase incorporated into the subsequent phases.

Select the right bot(s) – It is very important to understand the benefits and constraints of the different bot types before one or more bots are selected. If multiple bots are selected, it is very important to ensure that the customer can seamlessly navigate from one bot to another.

Conversational AI platforms come in all shapes and sizes. Some are nothing more than point solutions to address specific needs. While speed and ease of development are paramount for businesses looking to gain a foothold in their AI strategy there are other aspects to consider too. If you’re a global company you’ll need language support and it’s important to note that while some vendors offer multiple languages, it frequently involves a whole new build to use them. The same goes for porting to different services or devices.

As the number of devices that users interact with every day grows from smart homes to in-car tech, enterprises will need to ensure that intelligent agents will work across them. Other enterprise features such as scaling, rollback, and collaborative working also have their pitfalls and challenges if not considered upfront. In addition, the future will be in these interfaces “talking” to one another so an open and flexible architecture will be essential.


Conversational AI platforms come in all shapes and sizes. Some are nothing more than point solutions to address specific needs. While speed and ease of development are paramount for businesses looking to gain a foothold in their AI strategy there are other aspects to consider too. If you’re a global company you’ll need language support and it’s important to note that while some vendors offer multiple languages, it frequently involves a whole new build to use them. The same goes for porting to different services or devices. As the number of devices that users interact with every day grows from smart homes to in-car tech, enterprises will need to ensure that intelligent agents will work across them. Other enterprise features such as scaling, rollback, and collaborative working also have their pitfalls and challenges if not considered upfront. In addition, the future will be in these interfaces “talking” to one another so an open and flexible architecture will be essential