Blog

Conversational AI market outlook: 2021 trends and predictions

January 19, 2021

Discover the latest trends driving the conversational AI market forward in 2021

The conversational AI market is continuing to mature at an accelerated rate as a direct result of the ongoing coronavirus pandemic. 2020 required many chatbot and virtual agent vendors to rethink their business models to meet client and consumer expectations. This led to an innovation boom - from self-learning AI to advances in voice technologies - the results of which will bear fruit over the next 12-18 months.

Alongside this technological growth, the conversational AI market itself has required that vendors adapt the way they develop and implement their solutions. The heydays of basic question-and-answer chatbots have made way for artificially intelligent virtual agents that not only offer instant, 24/7 assistance but also perform advanced transactions on behalf of customers.

Advances in conversational AI have made consumer self-service far more efficient and convenient than previously possible. And, as such, the expectations of what a conversational AI solution can offer, both businesses and consumers, have shifted. Companies no longer require one-size-fits-all chatbots and, instead, expect to deploy enterprise-ready solutions that take full advantage of the technology.

Here are four important market trends that will drive the business value of conversational AI well into the future:

1) Going 'chat-first' will deliver the fastest return on investment

Gartner says that the future of self-service will be powered by customer-led automation. By 2030, Gartner analysts predict that a billion service tickets will be raised automatically by chatbots and virtual agents - or their near-future cousins.

This makes a lot of sense. Chat-based self-service is a low-cost, low-barrier way to automate customer interactions at scale. And, as consumers become more accustomed to it, we will begin to see businesses capitalize on this trend.

Adopting a 'chat-first' strategy - where a company funnels all customer service traffic through a conversational AI solution - will enable businesses to play to the strengths of automaton by reducing support costs and driving up CSAT scores.

We have already seen this approach yield impressive results within the banking sector:

  • DNB - Norway's largest bank makes its virtual agent the first point of contact for customers visiting its website. In 2020, the virtual agent automated over 10,000 inquiries per day and accounted for over 20% of all customer service traffic across all channels.
  • Sparebank 1 SR-Bank - This medium-sized Norwegian bank currently automates 42% of total B2B and B2C support traffic thanks to its 'chat-first' approach. 75% of customers also report that they prefer to get help from the bank's virtual agent - even when given the option to speak with a human.

2) Virtual agent deployment timelines will be significantly reduced

If the pandemic has taught us anything, it's that many businesses were woefully underprepared for the unexpected surges in customer service traffic that resulted overnight. Those companies with a virtual agent already in place were able to mitigate significant spikes in inquiry volume, provided their conversational AI solution was robust enough to handle it. Others, however, were not so lucky. Many businesses were caught off guard and had to scramble to quickly build and deploy chatbots that were either not feature-complete or took significant time and resources to implement.

In 2020, we will see vendors find ways to reduce the barrier to entry for conversational AI solutions. Sandboxes-as-sales-tactics will be replaced by accelerated proof of concepts that leverage technological advances like self-learning AI to make the transition from legacy customer service channels to customer service automation faster and easier.

Responsibility will also fall on vendors to prove that their solutions can deliver on promises of a genuine return on investment from day zero.

Key questions to keep in mind when determining if a conversational AI vendor is the right fit, include:

  • Does the solution include scaleable Natural Language Understanding (NLU) that can process thousands of user intents (topics) simultaneously?
  • Can self-learning AI be used to bypass a 'cold-start' and assist with the development and maintenance of a virtual agent?
  • How quickly can an AI chatbot project go from development to go-live? Is it only a few weeks (desirable) or many months (undesirable)?

3) Data-driven chatbot design is more important than ever

Gartner predicts that by 2022, 70% of white-collar workers will interact with conversational AI on a daily basis. If those interactions are to be meaningful, conversational AI vendors will need to step up their game, making it necessary to move beyond the basic design principles that chatbots have relied on for years.

In order to have a marked impact on how customers perceive virtual agents as a useful tool, it will no longer be enough for vendors to provide market-leading technology. Using evidence-based design to inform various aspects of virtual agent development - from its personality and avatar to its visibility on a website - will be hugely important going forward.

The solutions that provide a deep analytics toolset, and other resources such as best practices, will be key to ensuring that companies can use conversational AI to its fullest potential.

4) The future of conversational AI is decidedly human

Much of the hype surrounding conversational AI centers around the black-box nature of the machine learning algorithms that power it. What's often left out, however, is that humans also play a big role in the success of a virtual agent or intelligent chatbot.

A 2019 report from Customer Contact Week found that 88% of customer experience professionals believe that artificial intelligence will improve and enhance their work, rather than replace them outright. This is contrary to the popularly held belief that chatbot technology and automation tools will take away human jobs.

Instead, experienced human agents working in customer service can be plucked out of call center support teams and upskilled into AI Trainer roles. These skilled workers can utilize their existing customer service experience to enhance chatbot development and improve the customer journey.

For those employees who remain in traditional customer satisfaction roles, the fact that a virtual agent can automate the bulk of a company's most common customer queries in real-time is an advantage. It will essentially change their day-to-day work, no longer necessitating the need for human operators to answer endless repetitive queries and resulting in work that is more meaningful and customer-centric than before.