By the end of 2021, Intelligent Assistants (IAs) of the automated variety will be everywhere. They will take the form of chatbots, voicebots, virtual assistants, conversational IVRs and personal virtual agents. They will be invoked over smartphones, smartspeakers, websites, voice consoles in cars and popular messaging platforms. They are finally emerging as “go-to” resources for frequently-invoked applications and services, often starting with a question to be answered and then progressing through search and, ultimately, successful resolution of an issues or completion of a purchase.

Triggered by the global pandemic, acceptance of Intelligent Assistants accelerated in mid-2020. It was a few short months in the making. Hundreds of millions of people were forced to “go digital” as they sought assistant from the banks, merchants, healthcare providers, airlines and government agencies that figured prominently in preparing for inevitable changes in plans. The leading firms employed “Conversational AI” to handle unprecedented volumes of calls or “chats”, each addressing novel, highly-personal issues. There was a spike of inbound voice communications because customers’ first instincts was to call a toll-free number to get real-time advice from a live agent.

The Moment of Truth

Entering 2020, only the largest companies were dabbling with proof-of-concept or pilot projects aimed at solving problems arising from incorporating “Conversational AI” into their customer care or employee productivity fabrics. Helpdesk here, FAQ there, everywhere a chatbot. When the rush of activity hit in mid-year, those businesses and government agencies moved many of these limited implementations to the critical path with customers and employees as a matter of expediency. That was the moment of truth.

Many brands and agencies were pleasantly surprised to find that higher levels of automation and “capture rates” coincided with increased customer satisfaction. That was a first! IAs provided an AI-infused conversational alternative to making multiple calls over a matter of days, only to be put on hold for hours at a time. Both contact center and digital service managers are adapting strategies that build on customer confidence in conversational IAs by training them to take on more tasks to benefit both customers and employees.

Made Possible through Commoditization of Cloud-based Resources

The new era is made possible by commoditization of core enabling technologies that include automated speech processing (ASP), natural language processing (NLP) and machine learning (ML). Collectively, they support such functions as intent recognition, translation and emotion detection. They live in the cloud and are accessed through application programming interfaces (APIs) from technology giants, especially Google, Amazon, IBM and Microsoft. But there are a growing list of others, and a slew of small, innovators that use commodities as raw material for solutions to known problems, including more accurate speech recognition/transcription, faster tagging of intents and greater security.

New capabilities are “exposed” and made available through APIs that are cleverly knit together by system integrators (SIs) or business process outsourcers (BPOs) that expect firms to spend tens of billions of dollars on what amounts to a “conversational layer” between customers, employees and IT resources. As a result, companies of all sizes are able to introduce new EIAs quickly and affordably.

Solutions For Non-Technical Developers

Departmental and non-technical executives are the subject matter experts (SMEs) responsible for bringing IAs to both customers and employees. Their efforts are bolstered by tools and platforms for IA creation, training, maintenance and measurement. All benefit from rapidly maturing technology from an expanded ecosystem of solution providers. Speech Processing, Natural Language Understanding, Machine Learning and even Cognition have become commodities accessed through APIs from Google, IBM, Amazon, Microsoft and IBM, among others, according to “free-to-fee” formulae that make it easier to justify with reduction in operating expenses and potential increase in revenue.

“Low code/No Code” development platforms reside “in the cloud”, where they connect with desired resources based on drag-and-drop menus that create Visio-like conversation paths. Nuance Mix is an example, but Nuance is not alone. Inference, now owned by Five9, offers similar resources that also include a multiplicity of pre-trained agents to speed implementation in select verticals. IBM, with Watson Assistant, brings a 60-day promise. Verint has a “Blueprint” to move projects from ideation through implementation.

Fulfilling CSA’s Promise

The result of all these developments is the realization of Conversational Service Automation’s (CSA’s) manifest destiny. It will not be automation for automation’s sake. Instead, the specific objectives and benefits are:

• Callers define their preferred paths to task completion: It starts when an IVR offers new prompts or options. The option to arrange for a callback and hang-up “without losing your place in the queue” was just the beginning. Now companies offer a slew of “voice-to-digital” options that leverage investment in conversational AI across time and multiple channels.
• Companies support asynchronous conversations over any device: Smartphones will be the most common point of origination, but expect the use of branded wake-up words to initiate or revive ongoing conversations to embrace smartspeakers in homes and hotel rooms, intelligent infotainment systems in cars and kiosks in commercial shopping venues. Voice assistants will routinely shepherd customers to complete their selected tasks, fulfilling on the promise of asynchronous, optichannel conversations as the preferred engagement model.
• Consistently correct conversational answers, actions and recommendations at scale: We’re already observing accelerated acceptance of digital channels for healthcare, hospitality, banking and retail. Success will breed success and you can expect accelerated adoption in 2021.

In effect, the digital world will make its transition from “either/or” to “both/and”… or should we say “either/and”? Individuals take control of their commercial conversations routinely. They don’t have to choose a particular app or know precisely what they are asking for. As we review submissions for our “Decision Makers’ Guide to Enterprise Intelligent Assistants”, perhaps the most impressive advancements have been in the area of dialog management. Several solutions recognize that customers may go “off-topic” or change intents in the course of a multi-turn conversation. It is no longer a parlor trick to be able to let a person say, “wait a minute. What about financing?” in the course of shopping for a new car.

The fact that such a query might cause a live agent to shut down one set of connections or forms to open a link to another system no longer baffles a finely tuned IA. It will be able to support multiple threads, turns and outcomes. In the process IAs transform us into better humans.

Categories: Intelligent Assistants