Chatbots
How to Scale and Improve Chatbot Experiences
With the increased adoption of conversational AI, many enterprises are now looking at how to improve chatbot experiences and scale them across the business. When they first started on their bot journey the issue of scaling wasn’t necessarily top of mind as the focus was on getting the first bot to market quickly. For many the true impact of their chatbot project only emerged when it was in production.
Then the journey of fine-tuning the bot, increasing its functionality, planning chatbots for different business cases, or growing bots at scale begins.
Before jumping into the issue of scaling bots, let’s take a brief look at the market for conversational AI and how adoption has recently taken off, triggered partly by the pandemic.
The Impact of the Pandemic and the Acceleration of Conversational AI
When Covid-19 hit in 2020, many industries saw increases of up to 250% in inbound customer calls or chats to their contact center as concerned and distressed consumers turned to their business providers for information, assurance, and other needs. At the same time, contact centers were forced to close their facilities during lockdowns and get their agents up and running from their home locations.
These two sudden impacts totally upended even the best business continuity strategies. Many contact centers looked to re-platform and move to cloud solutions that enabled their customer service agents to work effectively while consumers pivoted to digital channels to have their needs met. Digital self-service became more than just a wish; it was imperative.
As Gartner forecast in their CIO Agenda 2021
“65% of CIOs saw an increase in the use of self-service by customers/citizens in 2020, 79% expect that usage to grow in 2021.”
To achieve the goal of advancing their self-service options for customers and employees, businesses turned to AI-powered automation solutions. In customer service and contact centers, management quickly focused on how they could deflect incoming calls and live chats to digital channels where a chatbot could respond to common queries and automate the tasks needed to fulfill the customer need. These contact deflection and automation solutions powered by conversational AI became key to containing as much as 75% of routine traffic into service desks. This video shows how a bot can alleviate pressure on agents and the associated costs of labor-intensive engagement channels.
As we look forward with anticipation to a post-Covid world, one fact that remains clear is that life as we knew it before the pandemic will probably be altered for good and many of these changes may actually be positive. Remote or hybrid working models have become the norm, even in industries that were reluctant to embrace the model. Companies have also learned the hard way that over-reliance on in-person interactions or the need to have a human worker present in an office to respond to customer requests or process transactions cannot be the only viable engagement model.
The impact became clear. Adoption of conversational AI solutions soared amongst companies with large volumes of customer and/or employee interactions as they honed in on their digital transformation initiatives and looked for chatbot solutions that could fully or partially automate key customer interactions and fulfill customer needs with little or no involvement from human agents. According to Gartner, penetration rates of Conversational AI increased by 20%-50% in 2020, compared to 5%-20% in 2019.
Five Ways to Improve Chatbot Experiences
Launching the first chatbot project is an exciting venture for most businesses but as they become more familiar with the technology and advance their implementations we see the following challenges or trends emerging.
- Creating Engaging Experiences beyond Simple Q&As
- Expanding Use Cases with New Features and Capabilities
- Scaling Deployments and Ensuring Resilience
In this blog, I’ll outline #1 i.e. how businesses can scale AI chatbots and create more engaging experiences:
Scale Conversational AI through Enhanced Chatbot Experiences
Many organizations have matured beyond their initial Q&A type bot deployment. Once they have a chatbot in production the natural next step is to improve on it and make the experience more personalized and human-like. Here are some examples of how a bot may be advanced in ways that improve the user experience. Expand the bot to:
1. Handle More Complex Tasks
What may start as a bot handling the top 8-10 most frequent customer queries may also expand to the bot handling more complex tasks. For example, a bot skill may be designed initially to provide a link to a web-page where the customer can download documents. Then it may be enhanced to the bot providing the customer with the specific document they request via document upload in the chat, removing much of the friction previously involved.
2. Handle Multi-Part Queries
A bot may be expanded to manage multi-part queries, such as “I’d like to return my item and get a refund to my credit card”. In this case, the customer is hitting two different workflows – the return of an item which may involve processing a return form and receiving a print label, and a refund of an amount to be paid to a certain account and confirmation of this on receipt of the return.
3. Increase Personalization
A big part of creating more engaging experiences is by introducing more advanced personalization. Integration of the bot with business systems allows customer data to be accessed securely and injecting the context into the chat. By doing so the bot has more history and information pertaining to the customer, thereby enriching the interaction, making it more meaningful, and enabling a greater degree of task fulfillment.
4. Weave Stories into Conversations
The concept of weaving stories into conversations is emerging as an important trend in user engagement. This goes beyond the simple question and answer format that doesn’t really represent the way we, as humans, have normal conversations. Stories represent the switches and turns that are characteristic of how we speak. So a consumer may start down a certain path asking a bot something very specific and suddenly change the context to something different. The bot has to maintain the context of the initial request while following the turn in the conversation. This is not as straightforward as it seems but advances are being made in this direction.
5. Be More Human-like
Then of course there is the concept of creating more human-like characteristics in how the bot responds and chats rather than very canned and stilted language. For example, a bot understanding sentiment in a customer’s utterance is an important step in managing a better experience, either through giving the appropriate emotional response or, in certain cases, even handing the customer off to a human agent. Then there is the whole field of creating a personality for the bot, even with a visual representation of a human-like, 3D avatar.
Summary of Enhancing Chatbot Experiences
Many chatbots commence their existence as simple bots which is the natural approach a business takes. As product managers, customer experience directors, and other roles monitor how customers engage with the bot continuous improvements are made. This can be by tuning the intents and utterances or by further advancing the bot functionality to create a better experience for the customer and boost the business value of the bot. We have witnessed this progression very much as a maturity journey.