Introduction

Conversational AI, encompassing digital assistants and chatbots, provides a multitude of benefits for both users and businesses.

From a user perspective, one of the primary advantages of conversational AI is the instant nature of the interaction. There are no waiting times, and the user gets an instant response 24/7 – if you think about customer service, this is a fantastic benefit.

There is no need to install a dedicated app, as typically conversational AI assistants are deployed in a wide variety of channels, most of them users already have access to. Think about social media (Messenger, Twitter) or communication apps (Whatsapp, Slack, MS Teams), etc.

With conversational AI, users can simply express their desires or needs, whether it involves scheduling time off from work or checking the status of an order, without the hassle of navigating through multiple web pages or menus. I think we have all experienced, at some point in time, the maze some websites present us when we try to get some sort of contact or support.

And to complement all of the above, the user uses their own words, on his own terms, whether writing or using voice – the power of NLP and AI will do all the heavy lifting in terms of “understanding“.

This seamless approach enhances the overall user experience by making interactions more intuitive and user-friendly.

If we look from a business’s point of view, conversational AI offers the opportunity to streamline customer service processes and to release the workload on call centers.

By automating responses to frequently asked questions, businesses can free up their staff to focus on more complex issues, all of that while delivering faster and more efficient service to their customers.

It’s easy to see the value added, especially during peak periods when call volumes rise. An automated agent in the form of a digital assistant ensures that customer demands are met promptly.

Conversational Sweet Spot

Identifying the sweet spot for each business is crucial and should be done early in the process.

Sometimes, the sweet spot may not be something that can be easily quantified. It could be an innovative initiative aimed at positioning the organization at the forefront of innovation, even if specific use cases are yet to be identified or defined.

However, there are several high-level categories that are commonly applicable across various organizations. Let’s now explore some of these categories.

Frequently asked questions

One of the most popular use case involves the management of frequently asked questions (FAQs), which is known for its simplicity and rapid implementation. These commonly asked questions typically already exist within the organization but are scattered across various sources, often lacking a proper centralized system for easy access and retrieval..

Implementing FAQs is typically a low complexity with rapid implementation

Businesses can use conversational AI to improve their employee’s and/or customers’ experience by providing answers to common questions about policies, expenses, and order status, just to name a few examples. This can be especially helpful for large companies with complex policies or product offerings.

High-frequency simple transactions

Another example is when it comes to basic, common (but highly frequent) transactions. These tasks include activities like submitting vacation time or giving feedback as an employee, which enhances the experience for employees. These are examples of internal transactions within an organization, but there are also examples of transactions aimed at the end-user or customer and also on those cases, this conversational automation will provide a smoother experience for customers.

High-load call center questions

When those high-frequency simple queries reach a customer support center, it has a bigger impact, as it will consume the time of a human agent. Conversational digital assistants can be valuable in providing support for call centers, particularly for common questions that have simple answers.

Take for example one of the following questions – “where is my order?“

In the act of checking an order status a human agent goes to the lengths of checking in the system the status of that particular order. There is really nothing much the agent can do when it comes to finding where that order or package actually is. This is a perfect example where the digital assistant could prove its value, by replacing the agent and doing the backend check on the order status. A simple use case, for which the usage of a digital assistant can provide an incredible improvement in the time allocated to agents.

Information gathering

Conversational AI can also be used to collect information from customers before connecting them with a human agent. This helps speed up the process at call centers, reducing waiting times and ensuring that agents have all the necessary details to provide the best service possible.

They can also be used to gather information before involving a human.

Can it solve every use case?

If the only tool you have is a hammer, then every problem looks like a nail.

Abraham Maslow

Not every business solution is best solved by a chatbot. As the saying goes, “If the only tool you have is a hammer, then every problem looks like a nail.” We need to remember the specific needs of each business and think carefully about which tools and technologies will work best to solve their particular challenges.

But not every use case is a conversational sweet spot.

While conversational AI has a lot of potential benefits, it’s important to recognize that not every business problem is best handled with a chatbot.

For example, tasks that require filling in a lot of fields, such as tax returns, might be better handled the old-fashioned way, with a form. Whether you use forms in your conversation or you choose to use a webpage to gather all the information, that is up for discussion, but ultimately using a pure conversational interaction to gather a lengthy and complex set of data, may not be the most efficient and pleasing experience.

Also, complex interactions that require a lot of back and forth, such as booking a flight or a hotel are not ideal for chatbots. This particular example resonates with my personal experience, as I typically use dozens of open tabs, checking different hotels, reviews, pictures, facilities, etc, and ultimately hunting for the best prices across various websites. These processes require a level of visual and comparative analysis that conversational AI alone cannot replicate or replace.

Similarly, engagements that aren’t really conversations, such as “command-line” tasks done by power users, don’t benefit from a conversational interface. This one is equally interesting, as I came to realize, users tend to be scarce with words. There is a natural tendency (which I too am guilty of) to employ single words or concise phrases to convey a message. This is not something we would do in a natural interaction with another human, but it’s quite common when we know there is a machine on the other side.

Shall we completely forget about these use cases?

Having looked into use cases that are clearly not a sweet spot for conversational AI, that does not mean mean conversations can’t play a part.

That being said, there are still use cases related to these scenarios where a conversational assistant can be incredibly valuable.

For example, a web application can incorporate an FAQ chatbot to assist users in navigating intricate tax forms. This chatbot would be available to address any queries users may have, eliminating the need to navigate through multiple pages or search for information elsewhere.

Likewise, when organizing a family vacation, a chatbot can offer suggestions on weather conditions, pricing, and other relevant details. In my personal experience, it could suggest a family-friendly hotel with excellent kid’s facilities, ensuring a peaceful and enjoyable stay by the pool.

Furthermore, even for experienced users who are proficient in command-line interfaces, there may be situations where leveraging voice commands through a conversational assistant proves to be more efficient. This integration of voice commands can streamline workflows, allowing users to accomplish tasks quickly and effortlessly.

Another example of a search where using NLP can give more accurate results. It also affords the opportunity to combine FAQ and search approaches. This is a topic I will address in a later chapter when I talk about “Strategies for Building an Intelligent FAQ Digital Assistant”

Let’s look at some more examples

Previously I talked about the specific example of “Where is my order”, but there are more examples in the domain of customer support, where we can easily identify value.

Customer support agents often encounter a high volume of requests from customers for refunds and exchanges. Companies typically have established policies in place to handle these types of inquiries. This means, that there is a workflow, or a sequence of checkboxes that need to be ticked, before the refund is approved. This is not up to the agent’s mood 🙂 So, the process of handling refunds and exchanges can become repetitive and monotonous due to the frequency of such requests.

“I want a refund”

This is where a digital assistant proves its worth by delivering business value.

Users can access the digital assistant without the need to wait or go through various channels. This fast-track access enables a faster resolution of their questions and eliminates potential frustrations associated with lengthy wait times or complicated navigation.

These benefits lead to higher customer satisfaction. The digital assistant provides quick and efficient responses, while skilled agents handle more complex issues. This creates a positive customer experience, meeting their needs effectively and fostering loyalty towards the business.

“Book next Friday as a holiday.”

Taking time off is a regular HR task that most employees undertake a few times a year. For such common and straightforward tasks, a digital assistant proves to be an ideal solution.

It works well conversationally; if you say “I want to go on vacation on the 29th for seven days,” the digital assistant can quickly book the vacation, taking into account weekends, holidays, and other factors.

Using a web app, you would need to find out when the last day is, and whether or not it’s a weekend or holiday. Overall, using a digital assistant can make the booking process a lot quicker and more straightforward.

Conclusion

In conclusion, it’s important to find the sweet spot for conversational AI and make the most of its advantages. While it may not be suitable for every situation, there are many cases where it can bring significant value. As technology progresses, we can expect to see more creative and effective ways to incorporate conversational AI into our daily lives.

References

[1] https://docs.oracle.com/en/cloud/paas/digital-assistant/learning.html

[2] https://www.colorado.edu/financialaid/forms/examples-tax-documents