
#contactcenterworld, @diradtech
Author: John Michne, Vice President, DiRAD Technologies Inc.
Today chatbots are being leveraged by companies for lead generation, customer service, placing orders, and so much more. Chatbots are an effective tool, but they are not magic. They are only as good as you make them. Unfortunately, not all chatbots are a success.
If you are considering developing a chatbot for your business, it is important to learn from past mistakes! Luckily for you we have compiled our top chatbot implementation mistakes along with tips for how to avoid them.
People are not interested in being misled. A recent experiment from Goldsmiths University and Mindshare concluded that most consumers did not mind communicating with a chatbot, but thought it would be creepy if a chatbot pretended to be human. Also, there are certain instances where people might prefer to talk to a human over a chatbot, or vice versa. For instance, if you need your password reset, interacting with a chatbot is sufficient. However, if you have a complicated issue and are concerned that a chatbot will misinterpret your unique situation then you would want to opt to talk to a real person.
The point is people want to know when they are talking to a real person and when they are not. Identifying your chatbot as a bot is pretty simple. You can take one of two routes:
Either method allows customers to easily decide if they would like to continue the conversation with the bot, or connect with a human.
The attraction to chatbots is the ability to easily interact with a company at any time. Customers don’t want to provide their full name, email address, phone number, home address, birthday, and favorite color (ok I am exaggerating, but you get the idea) before the conversation even starts.
Rather than having the customer fill out a lengthy contact form at the beginning, ask for the bare minimum to get things started. Once the chat is initiated you can ask for additional information in a more conversational way. Just be sure your chatbot is identifying why they need this information. For instance, "can I have your phone number in case we become disconnected?" or "please provide your phone number so I can look up your account information".
No one is perfect, and your chatbots aren’t either. At some point, someone is going to say something your chatbot doesn’t understand. Let’s consider this example:
Someone asks your chatbot "Do you have the shoes Jennifer Aniston wore on the red carpet in stock?"
Your chatbot has no idea what shoes Jennifer wore so it responds with, "Sorry I didn’t get that. Please try again."
The user then rephrases the question: "Do you have the heels Jennifer Aniston wore to the Grammys in stock?"
Your chatbot still doesn’t understand and responds with, "Sorry I didn’t get that. Please try again."
After three or four rounds of the user trying to rephrase the question to get an answer, they are officially deep inside the chatbot abyss with no way out but one- exit the chat. This is why it is so important when developing your chatbot to prepare for those inevitable, unknown answers so you avoid the oh so frustrating dead-ends.
How should you handle these errors?
First, it is ideal to have some variations of your error handling messages. No one wants to get the same error message over and over again.
Secondly, make sure your chatbot clarifies the issue. What specifically isn’t it understanding? What information does it need? In the example above the chatbot could have said, "Sorry, I am not sure, could you tell me the brand you are looking for?" and continue the conversation that way.
Finally, if all else fails, you should give the user an out. If your chatbot cannot answer the question or handle the request, just say so and give them the option to connect with someone who can help.
Before going live with your chatbot it is important to run testing that simulates live customers. However, testing and fine-tuning should really be an on-going effort well after the chatbot goes live. Leverage chatbot analytics and user testing to improve the chatbot experience. Make a note of unanticipated statements and add them to the response engine. It is important to continuously monitor users interact with the bot, which lines are well-received, which ones did not, and where people got stuck.
There are no shortcuts when it comes to developing a chatbot– you’re basically creating and training an "entity" to interact with your customers. Every possible iteration of that interaction needs to be considered. Bots might be smart, but the creator has to put in the work to make them that way. There is no magic in the backend, even with "artificial intelligence".
The way chatbots are developed has evolved in a short period of time. These advancements allow for chatbots to better understand natural language and be utilized for more advanced functions. If you are relying on old technology, you are living in the past and will not keep up with current customer expectations. Plus, you will severely limit your chatbot’s future capabilities.
It is important to stay up to date on chatbot trends and use emerging standards from companies like Google and Microsoft. The chatbot technology you choose should allow you to use artificial intelligence on the backend and recognize natural language for the best results.
Of course, it does! There are so many chatbot options out there, and not all are created equal. Remember, a poorly designed chatbot is not an asset, it’s a liability. And a rather expensive one at that. Interested in talking with a chatbot development expert?
About DiRAD Technologies:DiRAD implements technology solutions including call center software, outsourced customer care, Interactive Voice Response (IVR), mass messaging, and AI-based, omnichannel chatbot & voicebot technology.
Published: Monday, July 5, 2021
4.) | Consilium Software Consilium UniAgent™ Today’s call centers need their agents to work more naturally, and respond better and faster. Here, we talk about the innovations Consilium has made with the latest desktop and mobile agent releases of the UniAgent™ CRM connector. From small enterprises to Fortune 500 companies and governments across 110 countries, UniAgent™ comes into action right at the moment of reckoning, making every contact center experience engaging and personalized, powered by simplified customer identification and anticipating customer needs. UniAgent™ provides CRM integrations between all the Cisco contact center solutions (UCCX, PCCE, UCCE and HCS-CC) and over 20 CRMs, such as Salesforce, Microsoft Dynamics 365... (read more) |
7.) | NovelVox NovelVox CTI Connector NextGen Cisco Finesse, Genesys & Avaya Embedded CTI Connectors for Salesforce, Microsoft Dynamics, ServiceNow, Zendesk & more. |
I am checking out all the amazing and daily updated content on ContactCenterWorld.com and networking with professionals worldwide
Send To Friends Post On My Wall