Can a Chatbot Help My Business?
The short answer is “Yes.” But you, as a business owner or manager, should form your own opinion as to why you should consider using chatbot technology for your business. The first step in forming a strong opinion is to figure out the value you want to create for your business or your customers. Will generating this value set have direct or indirect monetary effects on your bottom line? Will it help reduce cost? Increase sales? Or are you after non-monetary values like improving customer experience?
Ask Yourself…
Part of forming a strong opinion and making an informed decision about this subject is asking yourself the following basic questions:
- Would it help my business if I had additional competent employees?
- Would extra workforce personnel enable me to expand my business?
- Wouldn’t it be wonderful if these employees only cost a fraction of my existing employees?
- Would it be helpful if these employees never got tired and stayed at work to help customers around the clock?
- Would it help my business if these employees were loyal, diligent, and disciplined?
- If they never pursued their own agenda and always followed directions?
- Would it help if my customers and prospects could receive help from employees anytime they needed without ever having to wait?
- Would it be helpful if my customers could interact with my business 24/7?
If you answered yes to most of the above questions, then a chatbot is right for your business. The good news is that the climate is ripe, since your customers have already accepted the concept of interacting with automated conversational technologies to get help, shop, and receive information. People would much rather get instant help from a robot than have to wait to get help from a human. The concept of conversational technologies promises assistance and value in real time, effortlessly, and via contextual engagement. It plays right into customer demands in this day and age.
Gartner predicted that in about two to three years, over 70% of all first-level business-to-consumer interactions will be handled by some form of AI and machine learning.
This is because people have not only embraced the benefits of new conversational technologies, but technology capabilities have also risen to new heights, helping raise a business’s bottom line by increasing business availability and bandwidth without having to hire more human resources.
The following figure shows why now is the right time to get involved in this trend:
The fact is, most of us have already been exposed to conversational platforms in our personal lives. We use smart speakers and virtual assistants on our phones to set reminders, execute simple tasks, and shop. These systems provide the level of convenience people love. And it only makes sense to have access to the same level of convenience when doing business.
Let’s take a facility rental business as an example, where people have to call and talk to an agent to ask questions about the items they are interested in renting. Customers would call in and wait for an agent to become available for a quick consultation to ask for what they need and receive the information that helps them to place an order. Here is a list of inefficiencies in this model:
- Customers cannot get help after hours and, without proper information, won’t place an order.
- Customers have to wait before they can talk to an agent.
- There is not a visual channel where the agent and the customer can exchange rich information like pictures, videos, etc.
- Customers are usually not comfortable providing sensitive information like credit card numbers over the phone.
- Placing orders over the phone is not convenient (spelling names, addresses, etc).
Now let’s consider the same scenario where customers interact with a chatbot instead:
- Customers can initiate contact just by typing what they need. They don’t have to wait, since the chatbot immediately answers and starts helping them no matter the time of day.
- Chatbots can provide rich information like pictures, videos, and comparisons over the same channel.
- Customers can place orders easily by typing or tapping their screens and through providing the required information with confidence.
- Customers receive important notifications about their orders in real time.
What’s remarkable is that a business can provide a much better customer experience at lower operation cost. Chatbots can eliminate operational inefficiencies, increase business capacity, and produce multifaceted values for business and customers.
Value For Customers
- Can receive help at any time on their own terms
- No wait time, can be served in real time
- Receive consistent user experience no matter the contact channel
- Effortless contextual engagement
- Rich information exchange experience
- Receive reminders and order notifications in real time
Chabots can not only help customers but also help company staff and relieve them from boring and repetitive tasks so they can focus more on their main duties.
If you analyze the activities of your workforce, you’ll find that they have to perform lots of undesirable, repetitive, and monotonous tasks that can be automated to free up their valuable time and energy. Virtual employee assistants, a form of chatbots, aim to address this issue by providing business automation to help relieve your workforce of tasks that waste valuable time and energy. Workers who have more time to think, bond with their peers, and work together are more satisfied and loyal to their employers and tend to go the extra mile to take care of customers.
Value For Business:
- Automate tasks and save time and energy
- Free up human resources to maximize their real potential
- Increased employee satisfaction and loyalty
- Extend business operation to 24/7
- Reduce operation cost
- Increase business bandwidth
- Improve customer service and customer experience
Successful Chatbot Deployment Checklist
Now that we’ve established the fact that chatbots can help any business transform to a better version of itself, let’s focus on what makes a chatbot project become successful:
I have to say that, first and foremost, the right attitude is key. Although the ultimate goal of deploying conversational technologies is financial gain, my suggestion is to first focus on creating and maximizing the value for your customers and your company. The perceived value will naturally lead to financial gains on solid grounds as the project matures over time. If the goal right off the bat is to increase sales numbers, I guarantee your chatbot project will not reach its maximum potential.
Once you pin down the value you want to generate, you must put together a list of metrics to gauge the performance and find areas that need improvement. This is very important. Without proper performance metrics defined, it’s not clear whether the chatbot is fulfilling promises or if it’s improving over time. Professional chatbot platforms provide the ability to customize performance metrics and scorecards to exhibit performance trajectory over time.
Commitment to Success:
Chatbot deployment is not a plug-and-play operation. Just like training a new employee and monitoring their performance over time to put them on the path to success, you must be committed to making your chabot a successful addition to your business. You need to closely monitor the performance, train, evaluate, and tweak to gain progressive results over time.
Start Small and Score Small Wins
Start small and create basic proof-of-concept projects and score small wins before you start thinking about creating a complicated chatbot with lots of skills. Every single skill requires a well thought out conversation flow, integration to other digital systems, and maybe even custom software development. Therefore, you don’t want to overwhelm yourself with a massive project. After successfully designing, deploying, and testing one tiny project, you can move on to the next skill and make your way towards creating a chatbot with multiple skills gradually.
For example, you may be unhappy about the fact that a large number of your website visitors remain anonymous and will not convert to paying customers. You may want to design a chatbot for your website to engage anonymous website visitors in interesting conversations and encourage them to open up and identify themselves while talking about their needs. This will help create relationships with website visitors who would otherwise stay anonymous. A chatbot with lead generation skills can help with this situation. Once you create a good conversation flow, you need to think about how and where to save the information the bot captures during the conversation. You may simply want to email the gathered information so it can be handled manually later, or you may want to integrate the chatbot to your CRM system to generate new leads automatically. Of course, the latter may need custom software development to automate the process. This is a relatively small project with a clear scope that can be built in a short period of time to help solve a very challenging problem.
Let Everyone Know…
Chatbots will not be helpful if people can’t find them or don’t know about them. Make sure all your customers know about your chatbot and how to take advantage of it. Advertise your chatbots in your phone greetings, on-hold messages, website, email signatures, or offer them when talking to customers over the phone. If customers know about your new capability, they will use it. Once again, it’s important to highlight the importance of integration between your different systems to achieve automation and provide a great customer experience. So, for example, after resuming business post COVID-19, lots of customers may call you on a daily basis to ask basic questions about new guidelines for using your services. You may not have enough staff members to keep up with the surge in the number of telephone calls and handle them promptly. You may think about creating a chatbot that can easily take these questions and properly respond to them to reduce the burden on your employees. The FAQ chatbot engages people in a text conversation and provides answers to the most frequently asked questions with no delay. You can let people know that they have the option to easily text you and learn about the new process. Simply change your phone greeting to announce the new capability and instruct them to start a conversation with your FAQ bot. If you have proper integration between your telephone lines and your texting solution, it’s possible to shift the conversation over text in real time where customers can get answers to their questions from your new chatbot in no time.
Support Various Messaging Channels and Meet Customers Where They Are
Chatbots thrive over messaging channels. This is great because people love texting as well and want to do pretty much everything with texting. Your chatbot project must support different messaging channels and provide the same consistent support over landline texting, website chat sessions, kiosks, social media, popular messaging apps, mobile apps, etc. Chatbot conversation logic will be the same; it just connects with customers via different channels to engage with them where they are.
It’s important to pick a chatbot platform that is channel agnostic; otherwise, it will be a one-trick pony like Facebook Messenger chatbots that cannot be used on your website or live chat bots that cannot answer Facebook messages.
It’s also important to understand the nuances and differences between different messaging channels. For example, you can click on pictures, scroll through carousels, and tap on clickable bubbles when messaging via Facebook Messenger, while these options are not available in plain old text messaging platforms. So you need to take into account the channel while building your conversation flow and present questions or information in the proper format accordingly. Sensitivity of the information exchanged is also very important. A banking bot might need to ask for sensitive personal information like bank account numbers, Social Security numbers, etc. It is very important to design the conversation flow properly to avoid exchanging sensitive information over channels that may not meet regulatory standards. In such cases, you would design your conversation differently between a non-secure channel, like texting, versus a more secure channel, like a website widget.
Integrate with Operational Systems and Achieve Task Automation
It’s still an upgrade from dead silence when it comes to greeting customers and answering their basic questions automatically when they contact you after hours. But it’s just the beginning for providing outstanding customer experiences that continue to impress. To expand the skillsets and functionality of your chatbot, connection to other operational systems such as CRM, ordering systems, ticketing systems, calendar, booking system, etc., is required. This will enable the chatbot to send commands to third party applications and execute actions like generating leads, sending emails, preparing reports, creating accounts, checking inventory, booking appointments, or whatever else it is trained to do.
Consult With Pros
Although it’s possible to use DIY tools to create a basic chatbot, you should consider partnering up with a professional team to help integrate your bot into different software applications/services to improve functionality. Not to mention the importance of security, access, and exchange of sensitive information. The experience of professionals in this field will help avoid putting your business or customers at risk.
Conversation Flow is the Heart and Soul of any Chatbot
Who isn’t impressed by a well-spoken individual who eloquently asks interesting questions and persuades the audience to open up and volunteer information? Conversation flow is the heart of a chatbot. Designing a conversation flow is both technical and artistic. The tone and the tenor of the chatbot should reflect the personality and the culture of your company. After all, it’s an extension of your business and should portray what you stand for. If a chatbot conversation is boring, misleading, or not easy to follow, people will stop talking to it and bounce. You should spend ample time thinking about how to maximize the value customers expect to receive from conversing with the bot and navigate the conversation toward the defined goals.
Handling Exceptions
During a conversation, customers and prospects might ask questions that the AI chatbot is not trained to handle. These are called exceptions. A good conversation flow gracefully handles exceptions to avoid disappointments and dissatisfactions. I have seen many glitchy bots that simply respond with “Sorry, I don’t know how to do that” or “Sorry, I didn’t understand that” to a very basic question just because it’s not trained to handle the question.
Imagine while at a hardware store, you spot a store attendant walking in the plumbing aisle and stop him to ask a question about your plumbing issue. The attendant happens to handle the electrical department and does not know the answer to your question. It would make you mad if he told you he doesn’t know the answer to your questions and just walks off. The right way to handle this exception is for them to explain that they work in a different department and, while they cannot help you personally, they will locate someone who can and stay with you until help arrives. Your chatbot must handle exceptions the same way. If the chatbot does not understand a question or is not trained to handle a query, it should immediately escalate the conversation and involve a human agent or handle the situation.
Human In The Loop
Just like live agents have access to supervisors and have the option to involve them on certain calls, chatbots must have the option to involve human agents when needed to handle customer expectations. There are multiple ways to do this. One way is for human agents to monitor exchanged messages in real time and intervene as soon as help is necessary. However, this method is expensive and resource consuming when many parallel conversations are in session.
Another method is to design the conversation flow with paths to human agents when exceptions occur. In other words, if the chatbot does not detect the intention of the customer or does not understand a query, it automatically follows a path to human agents while letting the customer know that the conversation is being transferred to a live agent.
Sentiment analysis and detection algorithms can also be considered for handling exceptions and spotting conversations that are not going well. Keep in mind that these algorithms are complex and expensive to acquire.
Metrics, Analytics, Reports
We started this article by highlighting the importance of forming your own opinion as to why a chatbot is necessary and how it can help your business and your customers. Once you pin down the value that you want to generate by deploying chatbots, you should immediately start thinking about the metrics that will help you gauge performance. Conversation flows that include skills where actions are taken on behalf of customers, like selling items, generating leads, booking appointments, etc., include chatbot duties.
To gauge the performance in this case, a metric that keeps track of the number of fulfilled duties as opposed to the total number of conversations will show the effectiveness of the conversation flow. If a chatbot fails to perform its duties frequently enough, the actual conversations need to be studied closely to find where the audiences bounce and stop following through. Tweak, re-deploy, and monitor to continue improving metrics.
Keeping track of metrics from the very beginning is an integral and essential part of committing to the success of your chatbot project. Consider using conversational platforms that allow you to customize and define your own metrics and produce analytics you need to improve on.
For instance, you may want your chatbot that provides first level customer service to create trouble tickets on behalf of customers. This is the chatbot’s duty. It makes sense to measure the number of tickets created against all conversations that were routed to the support conversation flow. If the number of created tickets verus support related conversions is relatively small, it means people who engaged with the chatbot didn’t follow through all the way to the end. This is clearly indicative of conversation flow problems that can be flagged by the reported analytics early on.
You need to continue watching the metrics, explore the conversations, and find areas of improvement to enhance customer experience over time.
Be Committed to Your Chatbot Success
Once again, unlike your average software application project where you design, test, and deploy it once and forget about it, conversational experiences are not plug-and-play and need continuous attention, nurturing, and improving. It’s all for good reason, too. Investing in artificial intelligence conversational platforms is like investing in a loyal employee who will stay and grow with your business forever and, therefore, deserves as much time and attention as needed to be trained meticulously and committed to its success.
Author:
Pejman Rajabian
Chief Technology Adviser
Zebyl Inc.
