It seems that the rapidly growing interest in the voice AI technology we’ve been observing recently is here to stay. According to the Marketsandmarkets report, the global conversational AI market is expected to grow at a CAGR of 22% during 2020–2025, reaching almost $14 billion by 2025.
So what are the forecasts and trends to watch in 2022? Let’s see.
Virtual agents with emotional intelligence
One of the main obstacles to why consumers can be reluctant to talk to voice bots is that they don’t have the ability to perceive and express emotions – in other words, they lack empathy. This has been true until recently. The new generation of intelligent voice agents is not only capable of discerning intonation, tone of voice, and mood of the customer but also can mimic human emotions themselves, allowing for a more natural conversation flow.
For example, NVIDIA has recently introduced the Expressive Speech Synthesis technology that can be used to generate human-like speech for multiple purposes. Another example is the conversational AI company Neuro.net that implements voice AI agents that are indistinguishable from humans in 99% of the cases.
Pervading corporate environment
The advancement of speech recognition and synthesis technologies has prompted more and more companies to adopt conversational AI solutions either for internal use or for external communications, or both.
According to Gartner, 70% of white-collar employees will be interacting with conversational platforms on a regular basis by 2022. Most CIOs that took part in the survey named chatbots as the primary AI-based tool used in their enterprises – and this trend is likely to continue. IDC analysts predict that by 2022, 30% of enterprises will use conversational AI solutions for customer engagement.
Assisting people with disabilities
Conversational AI is already being used to assist people with visual impairments, paralyzed patients, or just elderly people who are not very comfortable using visual interfaces. However, these applications are not so widespread yet, and there are certain issues arising, such as less effective speech recognition for people with speech impediments.
Further advancement of voice AI technology and involving people with disabilities into the development and implementation processes can drive the adoption of conversational AI as an accessibility tool. Considering the fact that such giants as Microsoft are now actively investing in such projects, we can expect more voice-tech products for accessibility purposes coming in 2022.
Better performance in noisy environments
One of the challenges with voice recognition is processing requests with massive background noise – especially when it comes to crowded environments. As a study by Deloitte shows, 11% of all patents in the voice AI industry are focused on improving the performance of voice assistants. These include solutions filtering out background noise and improving speech recognition quality.
Another challenge is refining the methods for activating voice assistants so that they are more accurate. For example, there’s a patent describing a method for lowering the likelihood of a voice assistant being mistakenly triggered by background noise. Hopefully, soon we can expect that voice assistants can be more widely used in busy environments, such as offices.
Today, most AI-powered bots are designed for specific application purposes – for example, guiding customers through the registration process, helping them solve technical issues, or informing them about forthcoming events or special offers. The more specific the task is, the more effectively and accurately it can be performed by virtual agents. The prospects of a general, or ‘strong’, AI capable of solving a wide range of issues still seem to be сloudy.
However, an approximation of a multi-purpose AI can come in the form of arrays, or assemblies of interconnected virtual agents, each specializing in a particular task. The ‘dispatcher’ bot at the entry of such an ecosystem can infer the client’s intent and then route the request to the corresponding virtual agent for processing. Deloitt’s study shows that nearly 7% of patents in the conversational AI domain concern multibot architecture.
Customer engagement through social media
Conversational AI has already begun to capture positions on social media platforms like Facebook, Instagram, and LinkedIn and is likely to continue this expansion as more and more businesses choose to automate their brand communication and marketing processes. In fact, using chatbots for customer engagement in social media has been one of the main drivers of the conversational AI market, which is expected to continue rapid growth in 2022, according to the Marketsandmarkets research.
Social media platforms, in their turn, encourage businesses to make use of the new technologies by enabling various chatbot integrations and offering built-in features. For example, Facebook has introduced Click-to-WhatsApp ads format that prompts instant communication with potential clients in WhatsApp which can be easily automated using the chatbot integration feature.
Growth of awareness about conversational AI solutions
However popular and omnipresent chatbots and other virtual agents may seem, the potential of conversational AI is far from being exhausted. There’re plenty of points of growth left geographically, technologically, industry-wise, etc. – in fact, the expansion has just begun.
For instance, despite the fact that Brazil and India are among the biggest consumers of smartphones, the awareness of the conversational AI solutions there is still less as compared to Europe or North America. But it’s only a matter of time that the new technologies will gain traction in these regions and boost the market growth.
Another example is voice bots that are expected to flourish in the next few years. According to the Statista’s forecast, by 2024, voice assistants will outnumber the Earth population, reaching 8.4 billion units.