2020 has seen an incredible shakeup in the conversational AI space, fueled by a global pandemic that has necessitated a shift in how vendors, businesses and consumers view the technology. That’s not to say that business has been bad. Markets and Markets reports that the conversational AI market capped off the year at $4.8 billion and is expected to continue that growth towards an estimated $13.9 billion valuation by 2025.
This sends a clear signal that confidence in conversational AI as a technology and a channel is continuing to grow. This growth can be attributed to the advantages that chatbots and virtual agents offer when it comes to delivering automated customer interactions at scale, particularly during a period where many people are discouraged from seeking in-person customer service.
As with any technology with a large amount of buzz behind it, however, it can be challenging to cut through the noise. On its Hype Cycle for Artificial Intelligence, Gartner has upgraded its prediction for the chatbot market to reach maturity within two to five years – declaring conversational AI as the leading use case for AI in enterprises today. While any business looking to implement a conversational AI project in the next 12-18 months should be wary of a crowded marketplace, there are several key trends that have emerged that will define the market and help separate those vendors with the right technology and approach as the market matures.
In 2021, there are three specific trends that, based on various reports and my discussions with analysts, clients and partners, will be responsible for adoption and growth. They are by no means the only drivers (other key trends are discussed at length here), but they help to give an overall view of where conversational AI is headed in the coming years.
COVID-19 has accelerated adoption and increased the demand for smarter bots
The International Data Corporation (IDC) published a report in April 2020 recognizing the widespread adoption of conversational AI. Businesses in virtually every industry, as well as government agencies and healthcare providers, used the technology as an effective tool to mitigate the enormous overnight increase in customer service inquiries that the pandemic created.
This increased rate of adoption had an interesting side effect: it quickly became clear that not all chatbots are created equal. While pre-pandemic, it may have been enough to launch a simple FAQ bot and call it a day, businesses like banks and insurance firms, with complex product and policy portfolios, were dealing with sometimes two or three times their daily traffic volume, and some solutions buckled under the pressure.
This means that in 2021 vendors need to step up their game, using proprietary natural language technologies to build smarter, more capable virtual agents that not only provided instant answers, but that can interpret the nuances of human conversation and act on key processes as well. A basic chatbot that answers questions on a few hundred topics is like building a house on a shaky foundation. It may look and work ok at first, but soon enough it will come tumbling down.
Scalable conversational AI solutions that can not only handle thousands of intents but also offer a broad scope, will separate the good vendors from the bad. Transactional capabilities such as automating loan forbearance and deep integrations with RPA, OCR and voice technologies will become the new normal for conversational AI, unleashing its fullest potential.
Shorter development cycles with technology to match
Another interesting conversational AI trend that has emerged over the past 12 months is the shift away from long, drawn-out pilot projects towards accelerated development cycles. Gartner notes that its client base, in particular, is wisely accelerating AI development as a direct result of the pandemic so that they can begin to see returns much sooner.
In order to keep up with this demand, the technology and implementation practices of vendors are already beginning to accommodate this shift. Self-learning AI will become a major force in 2021, allowing for a significant reduction in the time it takes to build and deploy virtual agents, while also assisting in their maintenance and improvement. The technology makes it possible to scan and index a company’s existing website and extract pertinent information to develop a useable model in a matter of hours – something that would normally take weeks and hundreds of man-hours to do manually.
With this acceleration in the development cycle afforded by self-learning AI, the barrier of entry to starting a conversational AI project will lower dramatically. Sandboxes-as-sales-tactics will be replaced by accelerated proof-of-concepts that will require vendors to prove that their solution can deliver on its promises and that businesses will start to see a genuine return on investment from day zero.
Data-driven virtual agent design will become critical
With more and more businesses relying on conversational AI to automate large portions of their customer service interactions, analyzing the data and applying the insights it reveals back into building better customer experiences will become key. By 2022, Gartner predicts that 70% of white-collar workers will interact with a virtual agent or chatbot in some capacity on a daily basis. Both vendors and businesses will need to, by necessity, move beyond the basic design principles that chatbots have relied on for years if they hope to effectively engage with their end-users.
We are increasingly seeing that everything from a virtual agent’s personality and avatar, to its placement on a company’s website, can have an impact on the customer experience. Market-leading technology will always be important but combining evidence-based design with a no-code or low-code platform that can be operated by existing customer service staff (and not a team of data scientists!) will become a critical factor in the success of conversational AI projects going forward.
With the many curveballs that 2020 has thrown at every market this year – not just our little piece of the AI pie – it will be interesting to see how the next 12 months shake out. What conversational AI has proven, however, is that it is clearly a technology that can be counted on others as being adaptable and capable of assisting businesses and consumers in the face of unprecedented challenges.