17 Chatbot Statistics that Speak Volumes of AI Advancement

Updated · May 20, 2023

Did you know that the first-ever crypto transaction happened when a guy paid 10,000 Bitcoins for two pizzas?

He didn’t want to cook, but he also didn’t want to go through the hassle of calling and ordering food himself. So, he posted his request on an online forum and offered to pay (what today is) an insane amount of money to whomever got him dinner delivered.

If only pizza chatbots had been available back in 2010…

By now, though, most of us have dealt with bots when we’ve needed assistance—at times successfully; at others, not so much. Our chatbot statistics will let you know which one of the two outcomes is actually more common.

Read along to find out how automated today’s conversations really are.

Key Chatbot Stats (Editor’s Choice)

  • In call centers, chatbot usage nearly tripled during the pandemic.
  • Chatbots could automate 73% of the administrative tasks related to healthcare.
  • Retail spending through chatbots will probably reach $142 billion by 2024.
  • Facebook Messenger alone hosts at least 300,000 chatbots.
  • 46% of customers believe chatbots are just roadblocks to real people.
  • In fact, the same amount of people claim they’d still prefer a human, even if a chatbot is faster.
  • But, at the end of the day, 48% of people don’t care whether a chatbot feels human-like as long as it solves their problem.
  • Chatbots’ conversations with people can increase revenues by up to 25%.
  • Four-fifths of people who have interacted with a chatbot rate the experience positively.

Chatbot Use Cases

The fact that chatbots can be quite helpful in customer service is a well-known truth now. But bots are actually good for much more than that.

Join us as we explore the chatbots’ vast world, starting from the first of their kind—a pseudo-psychotherapeutic machine.

1. Joseph Weizenbaum created the first chatbot in 1966.

(Source: Onlim)

The history of chatbots begins with one named ELIZA, created by MIT professor Joseph Weizenbaum in 1966.

ELIZA operated in a much simpler way than modern-day chatbots. For one, it didn’t use any artificial intelligence. Instead, it analyzed the speaker’s words for keywords it knew and then chose an appropriate response from a number of scripted options.

Although ELIZA wasn’t able to “learn” new things through interaction with users (unless its script was manually edited, of course), many were convinced that it possessed intellect. Mr Weizenbaum did design it to speak like a psychotherapist, after all.

In reality, despite being one of the most famous chatbots of all time, ELIZA was also the first—and, thus, much more rudimentary.

2. In a matter of days, Meta reduced the likelihood of its chatbot saying offensive things by 90%.

(Source: Bloomberg)

Do you remember Microsoft’s chatbot Tay?

That’s probably because the tech giant shut it down merely 16 hours after release due to the extremely offensive posts it started to make on Twitter. The company launched multiple iterations of its bot, all meeting a similar fate.

Meta’s been doing the same for some time. This month, it released BlenderBot 3, and, lo and behold, despite its mighty 175B-parameters-strong chatbot intelligence, Meta’s offspring started spewing anti-semitic conspiracies right away; it also claimed Trump was still president and even praised Adolf Hitler.

However, instead of taking it down, Meta decided to add a disclaimer box stating that it’s for “entertainment only” and that it may make “offensive statements,” and vowed to keep on improving it.

So far, this strategy seems to be working—the bot is now 90% less offensive.

3. During the pandemic, Chatbot usage in call centers nearly tripled.

(Source: Bloomberg)

Prior to the pandemic, chatbot usage sat at about 10%. Since COVID19 struck, it has risen nearly threefold, with call centers choosing to implement chatbots and other similar AI solutions 25%-35% of the time.

Estimations suggest this development could have otherwise taken up to half a decade to occur naturally. This is undoubtedly a positive turn of events for the chatbot industry and call centers, but not everybody is happy about it—workers don’t appreciate being replaced by robots.

Fun fact: Projections suggest that AI will render 400-800 million jobs obsolete by 2030, forcing 375 million people to switch occupations. Fortunately, artificial intelligence will create jobs, too—about 9% of the working positions in the US in just a few years will be AI-related.

4. Only 10% of people agree with the government using chatbots.

(Source: Invesp)

There are many companies using chatbots—some more successfully than others. But one entity most people agree shouldn’t rely on them is the government.

On the whole, it appears that the more “serious” and “sensitive” an industry is (in terms of the handling of personal data), the less likely the citizenry is to support the automatization of customer support. For instance, only 20% of people would be okay with it in banking, 15% in insurance, and a meager 10% in governmental institutions.

For comparison, 34% of customers don’t mind chatbots in online retail—which is where they’re most common anyway.

5. You can order a pizza from Domino’s by sending them an emoji…of a pizza.

(Source: Tatango)

Amid the many chatbot services available nowadays, we are particularly fond of Domino’s genius idea to automate pizza orders. All you need to do is to text them an emoji of a pizza—or just tweet the emoji—and wait till they deliver your food.

How do they know what you want?

Well, you can set it up on Domino’s website under “Easy Order,” where you can specify what your favorite pizza (and address) is. Then, with a single emoji, you can have it anytime you want.

(They still charge you, though.)

6. Chatbots could automate up to 73% of administrative tasks in healthcare.

(Source: Insider Intelligence)

The benefits a chatbot can bring to the table are many, and they’re particularly meaningful in the healthcare industry, where matters are often of life or death.

Suppose healthcare providers are able to make full use of them. In that case, they could automate up to 73% of administrative tasks, which would, in turn, open up resources to allocate towards more significant tasks.

All in all, estimates suggest chatbots can help the healthcare, banking, and retail sectors save $11 billion a year.

Chatbot Market

How can chatbots help increase revenue and profits?

The obvious answer is by automating customer support and allowing businesses to save on payrolls—but there are other ways, too. Adequate use of chatbots in retail can boost client acquisition and even convince people to spend more.

Read on to see how.

7. Projections suggest the market will reach $1.9 billion by 2027.

(Source: Fortune Business Insights)

With the latest advances in AI technology, chatbots have become sophisticated enough to start replacing humans in certain customer support roles. Accordingly, the chatbot market size is growing—and quickly.

It was valued at just $396 million in 2019, but analysts expect it to grow at a CAGR (compound annual growth rate) of 22.5%. This means it should reach $1.9 billion by 2027, quintupling in size in under a decade.

Fun fact: North America accounts for about a third of the global market.

8. Facebook Messenger alone hosts at least 300,000 chatbots.

(Source: Venture Beat)

One of the easiest ways to employ a chatbot for business purposes is through Facebook Messenger. Since 2016, the social media giant has offered the option to create chatbots integrated with Messenger, allowing businesses to communicate with their clients seamlessly.

After reporting an initial 33,000 bots, Meta (then still known as Facebook) updated that figure to 100,000 in 2017. Since then, the number has risen to at least 300,000.

Fun fact: Every month, businesses and customers exchange around 20 billion texts on Facebook Messenger.

9. 19% of all businesses use chatbots.

(Source: Tidio)

Even as chatbot technology has improved thanks to various AI solutions, far from everyone is keen to use it.

At present, just 19% of businesses have invested in a chatbot. Moreover, it’s primarily smaller firms that have done so—22% of them have adopted one, compared to just 12% of large companies.

One possible explanation is that well-known businesses don’t want to risk using a bot that’s not good enough and that could alienate customers. Today, chatbot statistics in 2022 show that 17% of big companies aren’t interested in chatbots at all, but the majority (71%) consider them as features to be added in the future (when they’ve gotten even better). 

10. Retail spending through chatbots will reach $142 billion by 2024.

(Source: Insider Intelligence)

An increasing number of people now make purchases through chatbots or other types of virtual assistants, including voice-based smart devices (like Alexa).

In 2019, such “chatbot sales” amounted to approximately $2.9 billion—a substantial figure, no doubt, but a tiny fraction of global retail spending ($26 trillion).

This is set to change, as projections suggest that spending through chatbots will rise all the way up to $142 billion by 2024.

11. Using chatbots for conversational commerce can increase revenues by up to 25%.

(Source: Medium)

There are many ways to use chatbots other than to automate customer support services. If you play your cards right, you could see substantial revenue growth.

Chatbot stats concerning “conversational commerce” show that businesses can increase their earnings by 7%-25%. How do they do that?

SnapEngage, for instance, comes with a proactive messaging function that targets qualified leads. This goes a long way when it comes to cart abandonment rates.

Fun fact: Nearly 70% of online shoppers abandon their carts—hence why companies are desperately trying to encourage people to finalize their purchases. Emails, chatbots, and shopping cart software can go a long way.

The Future of Chatbots

Chatbots have improved so much that it’s already hard to tell if you’re speaking to a human or not right away. But good technology alone may prove to be insufficient—many oppose the replacement of humans in customer service for reasons other than efficiency.

Let’s see what the source of customers’ reticence is.

12. Five messages are all it takes for 73% of chatbot conversations.

(Source: Tidio)

The average online chatbot is remarkably good at holding a conversation—in its own area of expertise. Sadly, though, it doesn’t get much of an opportunity to show off its talents.

The vast majority of conversations (73%) involving a chatbot end after no more than five messages are exchanged. In fact, fewer than 10% last for longer than 10 messages.

In other words, the average length of a chatbot-to-human conversation is 5.7 messages.

We at Web Tribunal can’t say for sure whether this means that the chatbots are efficient or that customers get frustrated and abandon the chat quickly. Maybe it’s a combination of both.

13. Almost half of the customers believe chatbots are simple roadblocks to real humans.

(Source: GetVoIP)

One of the most significant challenges with chatbot proliferation today has to do with people’s perception of bots as little more than roadblocks, obstructing access to an actual human.

In a way, it’s a perfectly reasonable concern—after all, companies use bots to handle simpler tasks and interactions, thus freeing up resources. But since many people will only contact customer service when their issue proves to be complicated enough to warrant asking for help, it’s understandable why 46% of them feel annoyed at the prospect of having to go through a bot first.

Alas, the customer service chatbot is by far the most popular type of virtual assistant, with 67% of all bots being used to help users when they reach out.

14. 80% of people who have interacted with a chatbot rate the experience positively.

(Source: Uberall)

Even though many appear to be skeptical of the technology, most people who have interacted with a bot had a good experience.

As it is, chatbot statistics show only 20% of users claim to be unhappy with the interaction—though, admittedly, a fifth of those were actually “very” unhappy.

Despite the positive feedback, nearly 60% of people in the US say they’re not feeling too enthusiastic about using chatbots more often in the future. In fact, 29% aren’t interested at all.

15. Some people (27%) aren’t sure whether their last customer support interaction was with a chatbot or a human.

(Source: PwC)

Can chatbots simulate people well enough?

Sometimes. Typically, it’s fairly obvious that a bot is, well, a bot. But sometimes, they’re so good that 27% of people couldn’t tell for sure whether or not they talked with a real human the last time they used a live chat solution.

This raises some interesting questions: Should businesses explicitly tell customers that they’re talking to a bot? Should they try to blur the line between the two further? Would that be ethical?

16. 48% of people don’t care whether a chatbot feels human-like as long as it solves their problem.

(Source: Business Insider)

Generally, people dislike AI that feels “cold” to speak to, which is why chatbots are becoming more human-like, boasting fun personalities and cracking jokes.

Or so we thought—a study found that 48% of users don’t actually care whether their virtual helpers are good interlocutors so long as they do a good job of helping solve their problem. Moreover, 19% said they don’t want human-like chatbots to begin with.

Nevertheless, it seems there’s a clear winner when it comes down to the human vs. chatbot competition. The majority of people would still go for human interaction, either because they believe a person would understand them better than a bot (60%) or simply because they prefer to speak with a member of their own species (56%).

Fun fact: Only 13% of people would be happy if bots replaced all human customer service.

17. Nearly half of all users would still prefer a human, even if a chatbot were faster.

(Source: Insider Intelligence)

People nowadays are generally appreciative of technology that simplifies specific tasks. Young people, in particular, seem to consider it a way to avoid unnecessary social interactions—but that doesn’t seem to hold true with business chatbots.

A study found that 46% of customers would rather speak to a human, even if that were the slower option. Bizarrely, young people are the most likely to avoid a chatbot—60% of Gen Zers and 50% of millennials prefer a live person when seeking customer support.

Wrap Up

There’s a significant dissonance between people’s expectations, experiences, and preferences when it comes to chatbots. The majority of us underestimate how smart bots have become and are pleasantly surprised upon interacting with them.

Yet, chatbot statistics reveal few would feel comfortable if a bot was actually all there was.

What can we say?

We, the people, are hard to please.

Jordan T. Prodanoff
Jordan T. Prodanoff

A wayfarer by heart, Jordan fancies journeying into foreign lands with a camera in hand almost as much as he enjoys roving the online world. He spends his time poking at letters and pixels, trying to transmogrify them into something cool.