We’re living in a time when interactions with robots are a normal occurrence. But it’s not exactly Rosie from The Jetsons (at least not yet). Most people correspond with digital assistants and chatbots on our smartphones and when we visit websites.
An SMS chatbot is just one form of bot that automates communications with two-way conversations. Chatbots can live in all sorts of places, from your website to WhatsApp or even inside an application. So, how do you know if an SMS chatbot would benefit your company?
To get some answers, we reached out to an expert. Joachim (J.J.) Jonkers got involved with chatbot development when the technology was on the rise. In this article, we’ll share some of his insights on how to use SMS bots effectively.
Meet one of Sinch’s chatbot masters
J.J. joined the startup Chatlayer when there were only around five or six employees. He worked as a Product Manager but wore many other hats, from marketing and support to web design. That’s how it goes at a startup.
The opportunity gave him firsthand experience with conversational AI - a solution that was quickly evolving and finding its niche in customer communications.
“It was an industry that was growing quite a lot. But it was also a space where we were trying out a lot of different things. The technology wasn’t at the level it is today. The whole story around where chatbots should live – your website, SMS, WhatsApp or somewhere else? How do you connect it to your contact center? We were figuring all these things out as we were building.”
- Joachim Jonkers, Director of Product – AI, Sinch
Sinch acquired Chatlayer in 2020, and J.J. has been along for the ride ever since. These days, he leads a team at Sinch that’s finding useful ways to integrate AI capabilities into our product offerings.
What are SMS bots?
Let’s get some basics out of the way. A chatbot is simply a piece of software that’s programmed to simulate one side of a conversation. There are both basic and complex ways of building chatbots, but they all serve the purpose of automating communications.
Conversations with an SMS chatbot take place via text messaging. That seems obvious, but it’s important to note that SMS chatbots have some limitations. For example, bots built for popular messaging apps or RCS messaging can deliver richer experiences with visuals, buttons, and interactivity. SMS bots, however, are limited to texts, which are also limited to 160 characters each.
Is SMS the right way to use a chatbot?
Short text messages may be exactly what your company needs, especially if you’re just getting started with chatbots. The advantage of SMS is its ubiquity and popularity among consumers.
New research from Sinch found that, when hearing from brands, SMS is the most preferred form of mobile messaging and comes in second only to email in most situations. Get the details when you explore The art and heart of customer connections.
J.J. agrees that one of the key advantages of SMS communication is how nearly every customer you need to reach uses text messaging.
“You want your chatbot to be where your audience is. If your audience uses SMS, then that’s where the bot should be. Any time you have a lot of customers who are using SMS as their primary communication channel, and you need to have some sort of interaction, chatbots help you scale those communications.” - Joachim Jonkers, Director of Product – AI, Sinch
Jonkers also says about 80% of the chatbot experience will be the same from channel to channel. Using the right partner and platform, if you have a web-based chatbot, it could very easily connect to SMS.
For example, with Sinch Conversation API, you could connect a chatbot to any channel we support. That’s ideal for building an omnichannel customer experience where a conversation with a brand may start as a website chat and continue via text messaging.
Common SMS chatbot use cases
Most chatbots have a defined purpose that’s built around a specific type of communication.
-
Transactional/operational: SMS bots can provide information and automated answers about purchases, account access, deliveries, appointments, and more.
-
Customer support: Chatbots are often programmed to answer frequently asked questions, which gives customers the help they need quickly and efficiently. More advanced bots could even help with troubleshooting.
-
Marketing: Chatbots can also interact with customers for promotional purposes. That could include anything from a conversation about how to claim an exclusive discount to entering a contest using an SMS keyword.
We asked J.J. to give us some specific examples of how SMS chatbots are used in the wild. He said one of the most common SMS chatbot use cases involves two-factor authentication (2FA).
Let’s say you get a text message with a code for accessing an online account. The problem is, you didn’t request the code, and that means a bad actor may be trying to access your account. An SMS chatbot could take things beyond alerting you of suspicious activity. The bot could also provide you with instructions on securing your account and even help you reset your password via text messaging.
Many businesses use texts for appointment reminders and confirmations/cancelations. With an SMS chatbot, you could take the next step and automate the process of rescheduling appointments when people need to cancel.
J.J. also told us about a major U.S. telecom company that used an SMS bot in the midst of the COVID-19 crisis. The Emergency Broadband Benefit program gave discounted internet access to low-income households during the pandemic. Of course, if you don’t have internet access, you can’t sign up online. The telecom’s chatbot was able to guide these customers through the application process using mobile devices and SMS.
How is AI used with SMS chatbots?
Artificially intelligent or “smart chatbots” use AI to enhance automated conversations. The most basic types of chatbots are rule-based or pattern-matching bots that deliver responses based on recognizable keywords or specific customer questions. More advanced bots use natural language processing (NLP) to form a deeper understanding of what customers are asking.
Different types of AI are also used in chatbot development:
-
Conversational AI: Unlike rule-based chatbots, those using conversational AI are built to go beyond triggered, pre-written responses. It mimics natural human interactions and follows a conversation flow.
-
Predictive AI: Predictive chatbots use data and analytics with NLP to provide a more personalized experience to individuals. This is what digital assistants such as Alexa and Siri do.
-
Generative AI: ChatGPT made generative AI a recognizable term. It involves a type of machine learning that generates new content based on the dataset from which it is trained. For gen AI chatbots, that means every customer interaction could be unique.
If there seems to be some crossover among conversational, predictive, and generative AI, you’re right. There’s plenty of overlap. In fact, J.J. explained that a chatbot could use all three types of artificial intelligence.
When your organization is just getting started with chatbots, a rule-based option could meet your needs while giving you more control over the flow of communication. But there’s also a world of possibilities to consider that can create engaging, automated experiences and personal connections through AI.
Sinch’s research on customer connections found around 30% of companies already use AI in conversational messaging. The survey also found 78% U.S. consumers say they’re most likely to interact with a chatbot for problem solving purposes. A combined 45% use chatbots to get questions answered while 33% have chatbot conversations when troubleshooting.
![Chatbot use cases chart from 2024 Sinch research](/sites/default/files/styles/x_large/public/image/2024-07/Chapter%204_Page%2029_%20Chatbot%20use%20cases%20chart.png?itok=2Gww-ur0)
Chart shows instances where U.S. consumers are most likely to interact with a chatbot, with 45% using one to get questions answered. Source: Sinch Customer Connections report, 2024.
Understanding the risks of chatbots and generative AI
While the possibilities are exciting, you’ve got to be careful about how generative AI is used in customer communication.
Most people who’ve played around with gen AI have seen how tools like ChatGPT and Gemini can provide incorrect and inaccurate information. Occasionally, artificial intelligence is also prone to hallucinations, which is when the use of large language models (LLMs) results in nonsensical outputs.
You don’t want your company making headlines for a bad bot. But that’s what happened to Air Canada after its chatbot gave a customer inaccurate details that cost him money. Then there’s the story of a major parcel delivery service chatbot going off the rails. It started swearing and calling itself useless, which is not a good look. The bot even wrote a poem criticizing the company.
Train your SMS chatbot
To prevent this type of unwanted AI behavior, those who build chatbots implement guardrails to keep them in line. Guardrails provide a framework that includes a set of guidelines your bot follows. This helps ensure it adheres to ethical standards and stays in compliance with specific laws.
You can also manage the experience by training the chatbot on a specific set of data that fits its defined use case. A chatbot meant for support, for example, could be trained on all your support documentation or product manuals. Chatbots could also be trained to provide answers based solely off the information in your company’s blog articles. It can even analyze customer service conversations from the past to better understand how to interact with humans.
Sinch’s Knowledge Base AI is designed to make training an FAQ bot fast and easy so you can take advantage of generative AI while keeping bots from veering off topic.
Know your limits
One of the clearest limitations of an SMS chatbot is the length of each message. Since a single text can be no longer than 160 characters, you can build a bot that delivers replies that fit within those limitations.
That could mean defining rules for a bot that generates its own answers, or making sure the expressions you write for the bot are 160 characters or less. As J.J. notes, if you send an SMS with 161 characters, the cost is twice as much.
In a way, the character limits of SMS can also be helpful. Text messaging is supposed to be friendly and conversational - but also short and to the point. No one wants to read a massive wall of text that a chatbot generates.
Another important consideration is deciding what kinds of questions your chatbot should not answer. In some situations, it may be more useful for a bot to point a customer in the right direction to find answers. For example, you may want people to visit a certain page on your website if they have pricing questions.
Even if it were feasible, it may not be the best idea to fully automate customer interactions with chatbots. J.J says there will always be situations in which a human is the most effective way to solve a problem. Let chatbots handle the more mundane and repetitive tasks.
“It’s helpful to look at SMS chatbots as entry level support representatives that know the answers to frequently asked questions and can solve some of the basic cases. But then, whenever you have a more complex problem or need professional expertise, it’s very common for the bot to transfer you to an agent.” - Joachim Jonkers, Director of Product – AI, Sinch
At a certain point, there are diminishing returns for programming a chatbot to solve complex problems. Typically, the rarer a situation or problem is, the harder it is to automate conversations around it. You may not want to invest time and resources into automated conversations that are tough to solve and rarely occur.
Using generative AI to build chatbots
While generative AI doesn’t have to power your SMS chatbot, it could be the perfect way to create an ideal experience for customers.
Chatbots are built using a practice known as conversation design. Part of this process involves mapping out the flow of expressions, which are the various phrases the bot is trained to understand.
For example, if you needed a bot that assists with pizza delivery, there are many ways to ask to order a pizza and many other questions and requests a customer might ask while ordering. It takes quite a bit of creativity to think of all the options, which is why AI can be an effective collaborator.
Generative AI can help you come up with the right expressions to include in your conversation design. That’s why Sinch is integrating AI-powered solutions into our platforms.
7 benefits of using an SMS bot
We’ve already given you some examples of how an SMS chatbot can enhance the customer experience, but let’s take a closer look at some reasons why it should be part of your communication strategy. Here are seven key benefits of SMS bots for your business and the people you serve.
1. Everyone uses SMS
Well, maybe not everyone, but pretty close. It’s estimated that there are around 7.2 billion smartphones in use around the world. For context, the global population is about 8 billion people. According to stats from SimpleTexting, 90% of consumers text every day and most check their messages at least 10 times a day.
SMS is the most ubiquitous form of messaging, and it makes no difference if recipients have iOS or Android devices. All this means an SMS bot can reach a very wide audience.
2. SMS is highly scalable
Because of its wide reach, SMS communications are an excellent way to grow a user base, increase sales, and nurture prospects. SMS is also one of the more affordable and cost-effective marketing channels in most markets. What’s more? With an SMS API, it’s fast and easy to integrate with your existing tech stack so you can start connecting with a global audience in no time.
3. People pay attention to their texts
Engagement rates for text messaging are mighty impressive. We’re talking 98% open rates as well as 36% click-through rates. Most people open nearly every SMS they receive, and they do so within the first few minutes it arrives.
4. SMS chatbots improve efficiency
When you automate text-based conversations through an SMS chatbot, you’re able to reduce the time, effort, and staffing it takes to respond to common customer questions. That means support staff can focus on adding more value and help solve more pressing, complex problems. Plus, bots never sleep. When used for support, an SMS bot is available to help 24/7.
5. SMS chatbots increase customer satisfaction
Launching an SMS bot could very well lead to happier customers. What else would you expect when their problems get solved faster?
According to MIT Technology Review, 90% of businesses it surveyed saw a measurable improvement in the speed of complaint resolution after implementing chatbots. Plus, 80% of respondents noted a measurable increase in customer satisfaction
By the way, SimpleTexting also found that using an SMS bot to collect feedback on customer satisfaction is one of the channel’s most popular use cases.
6. SMS bots can be multilingual
Whether you need solutions for a global enterprise or plan to expand into a new market, SMS chatbots can easily be localized for different languages and regions. Check out the long list of languages that Chatlayer supports.
7. Text messages are easy to personalize
Text messaging is already a very personal communication channel. So, it makes sense that conversational AI would be a perfect fit. When an SMS chatbot converses with recipients using a customer’ first name, your brand creates a more personal connection. Of course, personalization can go much further than names. Find out more in our article on personalized customer experiences.
Why stop with an SMS chatbot?
What if you could connect with customers on all the most important communication channels? What if those conversations were seamless and cohesive across everything from email and SMS to voice and mobile messaging?
That’s how Sinch is leading the way with our Customer Communications Cloud. You’ll find everything you need to make meaningful connections in a constantly evolving digital landscape. Plus, thanks to people like Joachim Jonkers and his team, Sinch is always innovating and improving our solutions.
When you're ready, contact us to learn about Sinch's conversational solutions, or check out our pay-as-you-go SMS pricing for details.