5 Questions to Ask When Defining Your Chatbot Strategy
7min read - iAdvize
It’s no surprise to learn that the global chatbot market is expanding rapidly, with chatbot adoption soaring in digital retail. Data from Grand View Research (GVR) found the market size at $430.9 million in 2020–and eCommerce dominates with a 24.30% market share.
What’s more, GVR projects a 24.9% compound annual growth rate (CAGR) through 2028. Other data from Fortune Business Insights identified a similar trajectory, placing the global chatbot market at $1,953.3 million by 2027.
So, chatbots are a massive force in retail–and they’re only going to gain even more traction. That means retailers need to take steps now to make the most of their chatbot deployments.
While rudimentary chatbot technology is widely available, it may not deliver enough power for today’s retail brands. Chatbots aren’t “set-it-and-forget-it” tools that you can implement without forethought.
If you’re implementing or upgrading chatbots for your brand, you need to take a step back and look at the big picture. Asking five questions can help you formulate a chatbot strategy that maximizes results for your business.
Question #1: What's the Purpose of a Chatbot Deployment?
You’ve probably heard the phrase “start with why” and may have even read the book of the same name. The principle behind Start with Why is best understood through the Golden Circle concept outlined in the book. In essence, the Golden Circle model works like this:
- The “why” at the center of the circle is the purpose or motivation behind actions
- The “how” are actions and processes that help accomplish the “why”
- The “what” is what you do–and how you prove the “why” to be true
The problem is it’s easy to start with the “what” when implementing any type of program. In the case of a chatbot initiative, the “what” is implementing chatbots. It’s something you can do without putting a lot of deep thought into why. But to be truly successful, you need to flip the script.
Asking yourself why you are implementing chatbots can help you define your purpose to guide you towards the right implementation approach. For example, the purpose behind your chatbot initiative may be:
- Delivering more expedient digital service
- Lowering cart abandonment and increasing conversions
- Deflecting inbound calls to your contact center to save costs
- Providing 24/7 coverage to greet customers online
The answers to those questions can help you define which types of chatbots you need. Also, it’s wise to create a long-term vision that describes where you are today and where you want to be in a few years. As with any program, your best approach may be to start small and plan to scale.
In general, chatbots fit into three categories, each with distinct characteristics and purposes:
Generally speaking, traditional chatbots are best for simple inquiries, like recovering a lost password. When your site visitors have more complex or sensitive needs, a conversational bot–especially one powered by advanced artificial intelligence (AI)–is going to be far more helpful.
You can have more than one type of bot on your site. To do so, you’ll need clear awareness of which pages are more likely to draw straightforward questions and where you may need a more advanced conversational approach.
Question #2: What Chatbot Solution Approach Meets Your Needs?
Research firm Gartner maps the primary styles of chatbots to three distinct solution approaches in what it calls a “sophistication continuum” for the chatbot market.
Gartner also points out that you’ll need to think through use cases for your chatbot when considering technologies. While some chatbot solutions are broad enough to support any use case, others focus on specific verticals or industries. Plus, other forms of chatbots target unique tasks, such as scheduling appointment or booking tickets.
What’s the right path for retailers? Most likely, chatbots with limited functionality won’t fulfill all your needs over the long term. You need an AI-powered bot that can take in the full spectrum of visitor queries, understand user intents, and take appropriate action. At times, that action may be handling the interaction completely. But it can also mean knowing when to hand off to an in-house or on-demand expert.
Question #3: How Will Your Chatbot Acquire Its Knowledge Base?
Every chatbot requires human intervention to acquire knowledge. A major difference between traditional and conversational AI chatbots is when and how these bots learn from their human counterparts.
With a traditional chatbot, people spend significant time upfront ensuring that the bot knows enough to function effectively. By contrast, conversational AIs learn continuously over time.
Also known as rules-based or transactional chatbots, these chatbots feature a playbook of rules established on the back-end of the interface. Often, they present users with a set of options and respond based on input. Each action by the user leads to predefined next steps.
Why does this occur? These bots rely on input from teams that program user questions and script anticipated responses. In setting up the bot, teams establish decision trees and keywords to cover expected queries from visitors. While new releases can expand the bot’s knowledge set, these bots don’t learn proactively based on user interactions.
Conversational AI chatbots are far more sophisticated than their rules-based counterparts. They use AI, machine learning, and natural language processing (NLP) to evolve their awareness of human language over time. Instead of depending on pre-determined, scripted outcomes, these bots seek to understand user intents first.
If you engage with a conversational AI chatbot, the bot will try to understand your intent upfront. Then, the bot delivers what it thinks is the best reply, based on existing data and past interactions. As the bot observes how people react to its responses over time, it learns and modifies how it engages on similar issues going forward.
Today’s most sophisticated conversational AI bots can be remarkably efficient to implement. Instead of having a long training period before conversational AI can be fully operational, you can deploy an intent library populated with thousands of user interactions fine-tuned to your industry. Once the conversational AI bot is in place, it can learn through user interactions and continuous coaching by your in-house and on-demand brand experts.
Question #4: Who Is Your Intended Audience?
Another thing to consider when formulating your chatbot strategy is the intended users. You’ll need to know whether you’re targeting current customers, pre-sales audiences that you want to convert–or a mix of both.
Chances are, you’ll want to give every person an opportunity to interact, at least over the long term. But it’s common for brands to target a segment of their site visitors at first.
To reach customers, you may want to position invitations to engage with a bot on select pages, such as your home page, support pages, or within customer account portals. For pre-sales audiences, providing chatbot access on product pages or during checkouts is an intelligent move.
Knowing your audience can also help you define an effective personality and voice for your chatbot. For example, if your primary market is teens, you won’t want a chatbot with the personality of a grumpy older man.
How can you get your bot’s personality right? One approach is creating a bot persona profile, similar to the marketing personas you create for your target audiences:
Source: Chatbots Magazine, The Ultimate Guide to Chatbot Personality
One example of a brand with a distinctive chatbot personality is Sideshow, a company that sells collectibles from popular films, comic books, and TV shows. When customers need support or have a question, they can engage with the friendly Sideshow droid chatbot. If the droid can’t help right away, he whisks visitors off to a Sideshow expert to help visitors complete their “quest.”
Question #5: Does Your Chatbot Connect Visitors to a Human When Needed?
Traditional bots can excel at addressing very simple questions. For example, you can ask a bot, “What is the weather today?”, and a just-the-facts reply of “High of 70 degrees and low of 55 degrees with a 10% chance of precipitation” is the perfect answer.
In this instance, you know exactly what you need. You can frame the question in a way that allows the bot to provide an accurate reply. Also, with this type of query, you’re looking for information, not advice or empathy.
When applied in a digital retail setting, these transactional bots can help site visitors who know what they are seeking. They just need help navigating where they need to go. For example, if someone is trying to locate a specific product or figure out how to contact customer support, a bot can provide links. Many of today’s transactional chatbots ask visitors to select from a narrow range of choices to direct users to the right destination.
All too often, however, visitor needs are too complex for a bot to handle. When that happens, users can get frustrated, especially if bots don’t interpret their concerns accurately or provide repetitive answers. As IBM points out, bots need to recognize their own limitations:
Perhaps the most human thing we can imbue into our bots is the knowledge of their own limitations. In other words, bots should know when to ask for help from a human agent when they’ve received a request that extends beyond their capability to respond satisfactorily.
In those situations, a seamless bot-to-human handoff is a must. Following a few best practices can help ensure that users receive the support they need:
- Identify when users are growing frustrated. Chatbots should be able to identify actions and phrases that can indicate that the customer may be hitting a frustration point. Indicators may be visitors using the same keyword phrase multiple times or using words like “annoyed,” “worthless,” or “stupid.”
- Acknowledge the bot’s limitations. If the chatbot has run out of options and can’t offer further assistance, let the user know. Ask the user if they would like to connect with a human agent.
- Don’t make users wait. Proactively offering live support is better than trying to deflect human interactions. When users do opt for support, connect them as quickly as possible–and be transparent about any wait times.
- Don’t make users repeat themselves. Capture the complete conversation between the bot and the user and pass it on to the human brand expert or support professional.
A scenario such as a malfunctioning product is not something a transactional bot can handle. A bot can be the first point of connection and provide troubleshooting tips from a knowledge base. But if those simple solutions don’t resolve the concern, handoff to a human is the next step.
Ultimately, while bots can be on the front lines of your website, they should never stand in the way of connecting with humans. As PwC research uncovered, 82% of consumers in the US say they will want to connect with real people as technology evolves. Yet, nearly 60% felt brands had lost sight of the human side of the customer experience.
Build the Right Chatbot Strategy for Your Business
It’s clear that you can’t just deploy an army of rules-based chatbots on your website and expect results. Today’s retail customers are digitally-savvy and know they have a world of information at their fingertips. They can conduct product research across multiple brands in minutes and don’t feel a need to stay loyal to any brand.
In this hyper-competitive environment, delivering high-quality customer experiences can make your brand stand out. A conversational approach that features advanced AI can help people who visit your site get the answers they need while still receiving the personalized attention they deserve. When the bot hits the limit of what it can provide, easy handoff to a human is essential.
The right technology and expert teams working together can empower you to achieve next-level service excellence. You can have the best of both worlds–world-class technology and passionate brand aficionados–and position your company as a trailblazer focused on humanizing the digital experience.
Ready to make your brand conversational? Our Conversational AI Playbook is a deep dive into fast-growing AI technology revolutionizing how brands and consumers connect.