Conversational AI Statistics: NLP Chatbots in 2020

Conversational AI Statistics

Conversational AI Statistics: NLP Chatbots in 2020

Being tech-savvy isn’t so rare anymore, not even for non-techies… Chatbots seem to be a functional and fairly common part of companies, big and small… Pretty futuristic, right? But let’s not get carried away just yet. Things are moving faster than ever, but where do we stand in the day-to-day reality? Are we really living “in the future”? What does that mean for your business? This in-depth compilation of the latest conversational AI statistics will try to answer just that!

Recent studies, surveys, forecasts, and other quantitative research into the progress of conversational AI (chatbots and voice assistants), highlighted a great deal of findings.

Despite still some reluctance and worries on the side of consumers and employees, chatbots seem to, little by little, win over the affection of majority. Furthermore, thanks to the lower and lower threshold of accessibility to AI tech, 2020 forecasts and trends are being dominated by the use of smart assistants.

Should you buy in on the trend? If so, how?

Let’s explore the AI bot landscape of 2020 to find out!


AI vs Conversational AI


First of all, let’s get some facts straight.

There is a difference between when people talk about AI in general and conversational AI.

AI covers every aspect of computing science that deals with “teaching” machines to think and act like humans do, and doesn’t necessarily involve language. Practical use of AI today includes, for instance:

  • Object and face recognition on images
  • Smart cars like Tesla
  • Google Maps analysis of the speed of traffic movements
  • Identifying the contextual meaning of emoji
  • Fraud protection & fraudulent transactions predictions
  • Credit decision and risk assessment in finance

Conversational AI is just one part of the AI universe. It works on the bases of combining machine learning with natural language processing (NLP) – the purely linguistic branch of AI.

NLP, besides serving chatbots and voice assistants, can be used in text prediction and grammar checking, sentiment analysis, automatic summarization, etc.


Adoption of Conversational AI Statistics: Leaders vs Laggers


So, how does the AI adoption landscape look like today?

According to the Salesforce study (State of Service 2018), the leading industry in the adoption of AI, in general, are financial services.

No surprise there! Banks and financial institutions have a lot to gain from fraud prediction and intelligent credit decisions that would take humans ages to make (if we even could process that much data).

However, more importantly, the leading industry in AI chatbot adoption is Media & Communication with 42%. Voice assistants are once again most adopted in financial services. In both of these categories, governmental agencies are the most behind with only 9% (AI bots) and 13% (Voice assistants) adoption

When looked at from the geographic perspective, India is taking the lead and strangely, the US is falling behind.

adoption-of-tools-and-technologies

When the study looked at the percentage of the organization actively looking for ways to use AI, tech companies, naturally, came out as leaders.

What’s really interesting, in terms of geography, India cemented its lead and the US kept its place as the lagger. However, such results always need to be regarded objectively. While in India a greater number of companies are actively involved in looking for ways to apply AI, in the US, it might be fewer but among them strong players like Apple and Google.

Another take on industry though! A different study (by Capgemini – 2019) focused on the top 100 performers in banking & insurance, retail and automotive. The result of their inquiry shows that almost 50% of banking and insurance leaders are already using smart assistants. Consumer products and retail leaders are not far behind as this industry was one of the earliest adopters of conversational AI assistants.

percent-of-organizations-deploying-chatbots

Regardless of industry or country, Adobe survey found out that 31% of top-performing businesses plan to implement AI within the next 12 months.

top-performers-planning-got-use-AI

And, of service organizations surveyed by Salesforce in 2019, 23% were already using AI bots while 32% claimed to do so in the next 18 months indicating 136% projected growth!

Planned-AI-use-among-service-organizations

So, what else is there to know about the businesses using AI chatbots today?

According to 2018 AI bot statistics by Relay, 58% of businesses using bots are B2B.

b2b-using-chatbot-tech


AI Acquisitions, Funding & Patents


To offer you a more balanced perspective, CB Insights (2019) shown that AI acquisitions have grown as the acquirers are becoming more and more diverse. Yet, despite buyer diversity and increase in acquisitions, US tech leaders are still the ones acquiring the largest number of AI companies and startups to Fastrack their AI efforts.

conversational-ai-stats-US-tech-giants-AI-acquistions

Also, while in the early 2010s the fight for leadership was head to head, today, Apple takes the stage.

tech-leaders-race-for-AI

On another note…

Lux Research (2019) follows the evolution of AI-related patents from 2014 to 2018. The statistical analysis counted the total number of AI-related patent applications in 5 key technology areas:

  • Fundamental AI
  • NLP
  • Voice recognition
  • Computer vision
  • AI processors and edge computing

AI-related-patents-2014-2018

While the US started off as a leader in applications for AI patents, 2015 was a turning point when China took over the leadership. Since then, China has been pushing fearlessly ahead.

It’s good to know where AI ideas are coming from. Still, the data tells us nothing about the quality of the patents or how it leverages AI. There are not stats that focus solely on patents in terms of conversational AI. Plus, a number of patents files is also conditioned by the country’s population size.

Still, the number of growing patents does seem to colorate closely with growing funding of AI-driven startups (Statista, 2019).

Ai-startup-funding

 


Conversational AI & NLP Market Predictions


As we have already mentioned above, NLP is a core part of conversational AI, however, it can be used in a plethora of other nonconversational scenarios. When assessing prediction for NLP market evolution in the coming years, the research from MARKETS AND MARKETS considered as part of the NLP market the following:

  • Information extraction
  • Machine translation
  • Automatic summarization
  • Text classification
  • Question answering
  • Sentiment analysis
  • Others (e.g., spam recognition, language identification, etc.)

They estimated that the NLP market size will grow from $10.2 billion in 2019 to $26.4 billion by 2024 reaching the Compound Annual Growth Rate (CAGR) of 21.0%. During the forecast period, the North American region is expected to account for the largest market size thanks to agile developments in infrastructure as well as the high adoption of digital tech.

NLP-market-prediction-2016-2023

The study concluded that the key factors responsible for the growth of the NLP market include:

  • Increase in the use of smart devices
  • Growth in the adoption of NLP-based applications and cloud-based solutions to improve customer service
  • Increase in tech investments in the healthcare industry

So much for NLP… But, that same institute also made a prediction prospect specific to the Conversational AI market.

The research predicted that the conversational AI market size should grow from $4.2 billion in 2019 to $15.7 billion by 2024 reaching CAGR of 30.2%, higher than the NLP market overall. The study highlights that major drivers market growth will include:

  • Increasing demand for AI-powered customer support
  • Omnichannel deployment
  • Reduced costs of chatbot development

conversational-AI-market-predictions

North America is, once again, expected to hold the largest market size while APAC is expected to grow at the highest CAGR. The study goes to explain that he highest growth rate in APAC can be attributed to the substantial investments made by private and public sectors to improve their AI and ML technologies which will result in an increased demand for conversational AI solutions.

One thing to take away from these AI & NLP statistics is the knowledge that AI is inadvertently on the rise. Hence, AI technologies such as chatbots are unlikely to just “go away”.

Let’s dive into HOW these businesses use chatbot technology in practice!


Chatbot Use Cases: Business Perspective


One of the most spread chatbot use cases is providing assistance and support pre, during and post-purchase.

Recent chatbot stats only confirm that.

According to Statista’s 2019 results, 78% of service organizations leverage conversational AI bots in simple self-service scenarios. Right in the second place with 77% is using bots to assess the type and difficulty of a query before passing it on to human agents.

Interestingly, over 70% of the companies also use bots to help not customers but agents to retrieve information and offer recommendations to resolve queries quicker.

service-organizations-AI-bot-use-cases-statistaThe widespread use of bots as part of support is not that surprising. The assistance chatbots can provide covers a myriad of business processes from internal to customer-oriented.

However, besides being helpful in the time of need, chatbot marketing use cases are becoming one of the hottest trends.

According to Adobe, in 2018, 28% of top performers were using AI to power their marketing.

top-performing-companies-use-AI-for-marketingIn 2019, 58% of Lanbot’s customer base used bots to generate leads and increase engagement. Customer support came in second with only 13.6% and product recommendation third with 8.2%.

2020 trends also begin to show new creative trends such as using chatbots in email marketing or as part of content marketing strategy.


Bot-Human Collaboration: Usage Statistics


Above, we already mentioned that one of the common uses of bots was using them in combination with human agents.

In 2019, Helpshift published its State of Customer Service Automation Report which provides a deep insight into bot-human collaboration.

Over the surveyed period of 6 months, monthly issues (web and handled only by chatbots increased by almost 20% (from handling 40% of issues to 59% of issues). While the total volume of customer issues increased by 24% from Q1 to Q2 of 2019 and bot interactions actually increased by an incredible 51%, agents saw a slight decrease in workload and bot+agent collaboration stayed relatively steady.

monthly-customer-issues-with-botWhat’s interesting, when averaged across those 6 months, inquiries on which agents and bots collaborated resulted in the highest Customer Satisfaction (CSAT) – 4.40 (on a scale 0-5). That is a 1% higher average CSAT than issues handled by humans only and 7% better than the CSAT best performing channel (messaging).

Furthermore, when the study looked at the total number of individual messages sent by all chatbots (bot outbound) increased 35% on average from Q1 to Q2. However, agent outbound messages increased only by 8% during that period. So, even as the volume of inquiries rose dramatically during busy periods, agents’ issue load didn’t change dramatically allowing them to maintain quality of work.

agent-vs-chatbot-messages-sent

In addition to improving CSAT, the bot-agent collaboration enabled an instant Time to First Response (TTFR). Agents were able to respond to urgent issues more quickly while bots took over the lower priority (more routine) inquiries.

This gave the agents the chance to respond to urgent and VIP queries more quickly and resolve complex problems more efficiently. Over the 6-month period, the agents saw:

  • 6% decrease in average TTFR from Q1 to Q2
  • 7% decrease in average time to resolve (TRR) from Q1 to Q2

The data shows that the implementation of conversational AI can improve overall customer service. Furthermore, it also allows organizations to achieve a consistent CSAT as well as better TTFR and TRR without significantly increasing agent headcount as the company grows.


A Case for Conversational AI Bots in Messaging


When talking about AI and NLP chatbots statistics and predictions in 2020, it’s impossible NOT to mention MESSAGING.

Instant messaging apps beat social networks when it comes to monthly active users in 2015.

In 2020 their power and influence are only getting bigger and Zendesk’s 2020 State of Messaging only confirms the trend. More and more businesses are using instant messaging as one of their communication channels. The top reasons for doing so include:

  • Faster TTR
  • 24/7 support
  • More personal interactions

The popularity of messaging is reflected in the above-discussed study by Helpshift.

When comparing TTFR and TTR across all communication channels during the examined period (Q1 and Q2 2019), the messaging channel came in as SECOND best.

TTFR-and-TTR-messaging

So… if it’s not the fastest, why all that hype?

Well, because despite offering only second-lowest average time to first response and time to resolution, the messaging channel scored the highest CSAT from all the channels – BY FAR:

average-CSAT-by-channel

Helpshift’s research attributes the popularity of and satisfaction with messaging channels to it’s the ability to host asynchronous conversations. That means the message exchange is not happening in real-time (it’s not synchronous).

For example, the users can message their queries and then go about their business, as usual, waiting for the response. If they did so on the web, they would have to wait for the agent to become available, regardless of the bot gathering the key information beforehand. Hence, even if the messaging channel is not the fastest to resolve their issue, it doesn’t feel as long as on the web. With messaging apps, users are not tied to the computer screen.

Whichever way you look at it, chatbots, though useful on both channels, produce a much stronger impact on messaging channels.

The message is loud and clear: Customers appreciate efficiency above all else when in pursuit of customer support. And, if bots can make that happen they are happy to interact with them.

If that’s not persuasive enough, according to Mediakix (2019), in 2020:

  • Messaging apps are expected to bring average revenue of over $15 per user
  • Messaging app revenue is expected to to be largely driven by chatbots

AI Chatbot Benefits for Businesses


The two previous sections on bot-human collaboration and messaging already mentioned some bottom-line-affecting benefits businesses can yield from chatbot implementation.

The aforementioned 2019 research by Capgemini highlights further advantages as per the three examined industries:

compnaies-yielding-high-benefits-from-AI-chatbots

Top recorded benefits include:

  • Reductions in CS costs
  • Daily savings on man-hours
  • Reduction in the number of CS calls
  • Increase engagement with digital assistants over time
  • Reduction in customer churn
  • Increase in handled interactions

It’s interesting to point out that the consumer products and retail companies had the highest success rate across all of these categories.

That same study also compared the benefits as per the companies’ state of organizational capability (organizational structure, – Employee awareness of conversational assistants, etc) and customer-centricity (transparency, personalization, skills, etc.) :

  • Leaders (high maturity on both organizational capability and customer-centricity)
  • Average
  • Spectators (low maturity on both organizational capability and customer-centricity)

chatbot-assistant-benefitsCompanies with advanced organizational structure and well-defined customer-centric strategy realized by far the highest benefits from conversational AI implementation.

The 18% of leaders saw the Net Promoter Score increase of 5 points; 30% of them achieved a 30% reduction in customer churn rate level; 37% saw an incredible 30% reduction in CS costs and 44% reported increased employee satisfaction.

This is a loud and clear message that just the tech is not enough. To get the most out of chatbots, you need to define & ensure:

  • Appropriate processes that encourage adoption
  • A clear strategy and goals you wish to achieve with conversational interfaces
  • Employee awareness of tech behind conversational assistants

There is no shortcut. If you want a conversational AI strategy to reach its full potential, you need to think it through. Simply slapping a bot on your website and hoping for the best is not going to cut it.


How Do Human Employees Feel About All This?


The stats above pretty bluntly show that the success of AI closely correlates with your having your employees on board.

A Salesforce study from 2018, discovered that 71% of customer service agents view AI as helpful to their job while 69% would like to learn more about the impact of AI on their job.

Only 27% were worried that AI could cost them their job.

human-employees-opinions-about-ai-bots

These results, once again, prompt companies adopting conversational AI to ensure that the employees understand how these technologies work and how they will impact their daily work routines.

More specifically, the agents interviewed in the Salesforce survey reported benefits such as improved organization; increased first contact resolution; reduction of handling time; increased CSAT and NPS, etc.

conversational-ai-chatbot-benefits

Overall, it’s safe to assume AI & NLP bots can be a huge asset to human agents, improving the organization as well as the quality of their work routines.


Customer Preferences in Business Communication


If you looked at the 2019 State of Service report by Salesforce below, you would think chatbots and messaging are from challenging traditional channels like email, phone calls, in-person meetings or live chat. Sure bots are not mentioned directly but customer preference for more traditional channels is clear.

preferred-channels-online-and-offline-by-generation

However!

Asking what people prefer is one thing, knowing what channel would actually be proved to be most useful in action is another. In other words, a customer might prefer to use email or live chat… but throw in an urgent question and waiting time into the mix and the preference is likely to change.

As the studies above showed, consumers are hungry for convenience.

And the reality is…

90% of consumers are more likely to do business with businesses with brands that answer inquiries immediately (LivePerson 2019). Immediate responses are not possible using only the traditional channels or solely relying on your human employees.

more-likely-to-do-business

Also, since messaging is really that convenient, 55 % of consumers are more likely to do business with a company if they could text them. Sending an instant message and not having to wait at the computer for the response saves customers a great deal of time. It also feels more immediate than having to send an email and checking inbox for a response.

more-likely-to-do-business

According to the 2019 State of Conversational Marketing report, people are getting more accustomed to the idea of chatbots. Comparing the answers from 2019 with responses in 2019, the preference to interact with people is getting lower (from 43% to 38%). Similarly, the number of consumers worried about a bot making a mistake decreased by 6%. And, most importantly only 14% of customers would prefer to fill out a form instead of interacting with a chatbot!

customer-more-acustommed-to-chatbots

In fact, when the Salesforce’s State of Connected Customer Report (2019) looked at customers’ opinions about conversational AI, the responses were strongly in its favor. Especially responses coming from the business buyers!

The inquiry shows that over 60% of customers are open to the use of AI to improve their experience. It’s also interesting to observe that when comparing end-consumers with business buyers, B2B customers are much more open and confident about AI. Though, the B2B clients’ openmindedness and trust in AI might be the result of greater exposure to AI as well as being more informed about it.

conversational-ai-customer-opinions

Do note that the weakest category in the survey is consumer satisfaction with the transparency of how AI is being used. Hence, if you want your clients to be more open and willing to try new technologies, educating them about the function of AI is a way to go.

Either way, it’s apparent, that the demand for personalized instant services comes with the willingness to try and adopt new channels and technologies.

After all, while customers still see live chat as the most convenient, chatbots are the winners at “Best in delivering 24/7 service” category!

chatbos-best-at-24-7-service

So, sure, when you ask your customers what they prefer, they will all say it’s the live chat, the human touch, personal attention, and instant responses.

However, in business, this is simply impossible.

If you want to give your clients the personalization and convenience they seek, you need to keep up with the tools and tech that allows it.


Chatbot Use Cases: Customer Perspective


Curious about how customers use chatbots that are already there, fighting for the AI acceptance cause?

The already mentioned 2019 Capgemini study asked consumers to what extent they used voice/chat assistants for a variety of activities from shopping, to playing music to checking their account balance.

Retail activities such as researching and buying products, creating shopping lists and checking order status were among the top use cases with 74%. Playing music, checking directions and making bookings came in third with 58%. And, financial activities such as funds transfers and bill payments came in third with 53%.

how-are-consumers-using-chat-assistants

More closely, the study looked at which type of assistant (chat or voice) did consumers prefer to use at particular stages of their customer journey today and in the future.

Regarding current preference, chatbots dominated activities such as personalized product recommendations, making payments, checking order status and a myriad of post-purchase services. In the future, consumers showed a great willingness to conduct most of the retail activities with voice assistants, provided the level of accuracy improves.

In retail,

current-and-future-preferences-chat-vs-voice

On the other hand, when it came to banking and insurance, chatbot dominated the current preferences throughout the whole customer journey. It also did better at maintaining relevance in terms of future preferences. This tendency might be caused by the more sensitive nature of the activities.

ai-stats-chatbot-ativities-and-future-preferences-banking-and-insurance

Either way, users are willing to use conversational AI assistants in a great variety of ways from the start to the end of the customer journey. It’s not just about customer support. Conversations can attract, engage and sell as well as they can provide support.

And while today people feel safer with using visual conversational interfaces in the form of chatbots, voice assistants seem to have a bright future.


Consumers vs Conversational AI: What to Keep in Mind


Communication preferences aside, it’s time to look at what you should expect to deliver if you indeed decide to go down the path of conversational AI.

1. Be Honest

First of all, no NLP or conversational AI is so evolved it can cheat a person into thinking it’s human. When using artificial intelligence, you better be upfront about it, people don’t like to feel like fools.

conversational ai statistics informed-that-talking-to-a-bot2. Be Fast

Secondly, a chatbot is appreciated for its 24/7 presence. Hence, it’s hardly surprising people expect them to respond as quickly as when they talk to a human face to face. 42% of surveyed consumers expect bots to respond instantly within the first 5 seconds and 36% expect them to do so within one hour. No other channel is held to such high expectations.

chatbot-expectations

3. Know Your Audience

Thirdly, the more digitally adapted your customer base is, the more likely they are to use and appreciate a chatbot. According to the State of Conversational Marketing report, the higher the number of owned connected devices by the customer the higher the likelihood he or she will trust your chatbot.

conversational ai chatbot statistics more-devices

4. Manage Expectations

Fourthly, the innovations you introduce raise the bar for engagement. The latest Salesforce service survey showed that innovative solutions and services change customer expectations and raise the bar or engagement.

59% and 77% of consumers and business buyers respectively said that offering messaging apps a communication and engagement channel affects their expectations and likely lowers tolerance for delays.

The situation is not much different for chatbots. 52% and 75% of consumers and business buyers respectively said that use of chatbots and voice assistant affect their expectation for the company.

factors-that-change-customer-expectations

Therefore, if you are willing to innovate, you must be willing to be held to a higher standard.

5. Enjoy The Success

But, fifthly and lastly, all that hassle around conversational AI and NLP bot is pretty worth it.

After having a positive experience with a chat/voice assistant; 72% of users placed higher trust in the compnay; 71% shared positive feedback with their close ones; and 64% gave the business high ratings or shared positive feedback on social media!

good-chatbot-experience-effects


Conversational AI Statistics: Conclusions


Chatbots are on the rise. Not only are people more willing to use them but the technologies are getting ever more accessible financially as well as from the technological point of view.

That doesn’t human conversations are losing on the importance nor that bots are ready to steal our jobs. AI & NLP chatbots CAN save us time and resources, though. More importantly, they can do it while significantly improving and personalizing the customer experience.

Better yet, thanks to the rise of affordable no-code conversational AI and NLP chatbot solutions, intelligent assistants are no longer exclusive to big corporations. Anyone can make their marketing conversational.

Don’t wait!

Start 2020 with building your own intelligent NLP chatbot for the web or create a clever bot for WhatsApp or Messenger… after all, messaging apps seem to be the best platforms to get the most out of your bot. Stats don’t lie!

👇 Register with Landbot for free today 👇

Barbora Jassova
barbora.jassova@landbot.io