Streamlining customer support and going beyond notifications with an SMS chatbot
Chatbots generated a lot of press coverage and attention over the last couple of years. It does make much sense from a business owner perspective – who would say no to potential automation of processes which currently require human agents? However, we rarely discuss the other side of the mirror. Do customers really appreciate facing a chatbot when they are in need of help?
Amidst the chatbots’ fanfare, it’s often hard to determine what use cases chatbots can already handle and those that would necessitate significant progress in their underlying technology (NLP – Natural Language Processing).
To get started, we will present the benefits to expect when implementing the chatbots and their requirements. Then, we’ll see how can chatbot be used for customer support and its alternative (IVR). Finally, we’ll explore a radically different use case for chatbots: proving services with text messages thanks to an SMS chatbot.
When do you need a chatbot?
Before we dive into the successful scenarios in which chatbots are used, let’s take a step back to consider how to determine whether or not a chatbot would be a good fit in a given situation.
Some necessary precautions are welcomed to avoid reaching the situation where an ill-conceived bot would solve a non-existent use case.
User Experience improvements
Successful chatbots are pleasant to interact with and have a real user base. It’s recommended to map the current scenario and compare it with a chatbot-powered situation to achieve such a result: it would help to precisely pinpoint the actual utility of the chatbot. Here are some good questions to help you get started on this process:
- How is the problem currently solved from a user perspective?
- Would the scenario be faster, more fluid, more accessible if a chatbot was powering it?
- Which channel would the chatbot be available on? Are your users there?
Resources currently expanded to solve the given problem
From a business perspective, chatbots are cost-reduction machines. Once you’re convinced your users would benefit from the chatbot (which provides it with potential for its adoption), you need to run the numbers on it. It’s also interesting to consider the human-bot work repartition; fully automatable problems are rare, the reality often mixes automated processes with manual human intervention when necessary. Here are some questions to help you figure this one out:
- What resources are you currently using to respond to the issue you are looking to automate?
- Would it make sense to allocate the time and resources necessary to develop a chatbot to automate it?
- What share of the workload currently assigned to the issue at hand would be handled by the bot? Can you estimate the workload left for human agents?
Customer Support Chatbots
Customer support lies among one of the most successful use case currently found for chatbots. They make a lot of sense both from the business’ and the user’s perspectives.
For the users, a customer support chatbot means that the support becomes available at any time, and answers almost instantly. Moreover, since the support is automated, it alleviates the fear of wasting someone else time’s with basic questions.
On the business side, using a chatbot for customer support means a reduced need for human agents. Depending on the implementation, the chatbot can also be used to tag incoming queries and sort them. It optimizes the time allocation of live agents and increases the precision of the analytics gathered by the support service. Finally, since the chatbot can be implemented on existing support channels (direct messaging channels or even the help center), it facilitates its adoption.
The base flow for customer support chatbots
You might wonder how a support powered by a chatbot looks like? Here is a simplified ticket flow:
- The customer opens a ticket: A customer facing an issue with the product or services is looking for assistance. He goes on the usual channel (the help center for instance) and initiates a conversation with the bot implemented there.
- Ticket tagging: Using NLP, the chatbot can recognize the main topic of his queries: the ticket is tagged with a high-level topic. (delivery, payment processing, website error, etc.)
- Immediate assistance: If the problem at hand is among one of your most recurring ones, the chatbot might have an answer at the ready. It can also direct the user to relevant resources available on your help center.
- Follow-up: With guided questions, the chatbot is then able to determine whether the issues have been solved with the resources he provided or if the user needs further assistance (->escalation).
- (Optional) Escalation: Before escalating to the ticket to a human agent, the chatbot can collect the materials required for the processing of the ticket – such as information about the user’s setup or a clear explanation of the issue.
- (Optional) Human-agent response: The live agent can then take it from here. Since the ticket is already tagged, the agent repartition can be optimized. Moreover, the agent will be able to access the conversation between the user and the bot to assess the issue quickly.
Customer support chatbots’ requirements and limitations
While the flow depicted above is fluid and can be effective if properly executed, it faces several requirements and limitations. Indeed, developing such a chatbot requires a large existing ticket base to analyze as well as high redundancy in the queries asked by customers.
Because of these restrictions, this kind of chatbots tends to be easier to implement in B2C businesses with large user bases. Besides, even if those two conditions are fulfilled, developing such a bot requires a significant budget and relevant technical profiles in your team for maintenance.
Automation for other businesses using IVR
Such a solution is hardly enforceable for a B2B company for instance because they tend to have a more limited user base as well as highly-customized clients setups. A chatbot wouldn’t be able to recognize the issues and then tag and handle them efficiently. However, it doesn’t mean that they have to forgo automation altogether. Other solutions are available to automate part of their support processes.
Many businesses are still handling their support requests or potential customer inquiries the old-fashioned way: their customers call them. To process the flow of incoming calls efficiently, they can implement a solution using IVR (Interactive Voice Response) and Automated Agents.
IVR let customers interact with your company lines using speech recognition or using their telephone keypad. The IVR systems will respond with pre-recorded audio to instruct the users on how to proceed. An IVR system is useful to collect information about your callers. It allows you to prioritize calls based on their value (such as reducing/eliminating the waiting time for high-value customers) and effectively route the incoming calls to the best-suited agents.
Efficiently sort through your incoming calls with IVRGet started with CALLR
Let’s finish our overview of chatbots and automation method with a very different kind of bots: SMS chatbots.
A closer look at SMS chatbots
We often picture chatbots in an OTT app setting (Messenger, Whatsapp, Telegram…) yet SMS chatbots come with serious advantages, the main one being their reach: SMS is one of the most common communication channels with over 5 billion users worldwide. Besides, while OTT apps require a smartphone, SMS on any phones even the most bare-bones ones. Here lies the critical benefit of SMS chatbots: they can bring smart services to the most basic phones.
TextEngine is one of the most compelling implementations of this idea. As the name would imply, it provides a search engine available through texts. It’s essentially a gateway to search the web without requiring an internet connection.
Going further with notifications thanks to an SMS chatbot
SMS are great to update your customer on their order status, such as delivery notifications. Indeed, SMS are read quickly (90% of SMS are opened in the first 5 minutes) but one-way SMS can get frustrating for the end user. With an SMS chatbot, you can enhance your user experience even further with simple workflows.
For instance, to increase the efficiency of deliveries, you can let your users interact with their delivery notification:
Your package XYZ will be delivered to your home at 400-468 Clinton Avenue, Brooklyn, NY 11238 tomorrow between 2 PM and 4 PM.
Answer “OK” to confirm the delivery or “Reschedule” if you can’t make it. You’ll be offered three alternative delivery times.
The bot can then suggest three alternative delivery windows and the user will confirm simply by answering 1, 2 or 3.
Such a chatbot is quite easy to implement. Indeed, the interactions with the SMS chatbot in the scenario depicted above are well framed. A similar process can be applied to other use cases such as appointment reminders.
Thanks to our partnership with Recast.AI, the chatbot building service, you can quickly build an SMS chatbot and integrate it with your existing business processes.
Bring your services to the masses with a SMS chatbotGet started with CALLR
Learn more on chatbots
If you want to dive deeper into chatbots, we’ve already covered the topic quite extensively. Our first publication on the subject covered the rise of chatbots in 2017 and the emerging use cases for SMS bots.
To get a better understanding of the underlying technology (NLP – Natural Language Processing), do read our layman’s overview – The Chatbot Masquerade. It explains the different steps of NLP as well as the challenges faced when one wants to design effective chatbots and the next challenges for their adoptions.