Path to conversational AI
POSTED : June 24, 2020
BY : Raja Roy

Most people touch their faces two to three thousand times a day. Being conscious of the surfaces we touch—especially germy phones, computers, keypads—and of the frequency we then rub our eyes and scratch our noses is now part of our everyday. As that awareness grows, so too will trends in technology that guide us more towards contactless experiences.

Conversational AI, voice interfaces and virtual assistants already live in our pockets, purses and on our wrists, aiding in the reduction of touches we make daily. And we’re witnessing an acceleration of their adoption as conversational AI becomes more human-like. By simply saying “OK Google” and asking your phone to make a reservation at your hair salon, Google’s Duplex will call the salon and interact with the person on the other end of the line. The AI is almost indistinguishable from a human.

The maturing of smart technologies has also expanded into the home, where ambient computing is contributing to more hands-free experiences. Google Nest Hub, with its ability to manage your thermostat, security cameras, robot vacuum, all using voice-based commands, combines AI and IoT to control the environment around us in a way that’s always-on.

In a contactless economy, consumers will expect more than just information-focused text chatbots. Businesses will need to incorporate more sophisticated conversational AI, like the technology developed by Google, into their customer experiences.

Conversational AI landscape

There’s a wide array of conversational AI-infused virtual agents in the marketplace. You’re probably most familiar with the simple FAQ chatbots on web pages, which are functionally narrow and often highly specialized conversational agents. Higher-level virtual assistants in a call center are capable conversational agents and can adapt to broader interactional natural language conversations of a transactional nature, rather than purely informational. Cutting-edge concierge robots built using a conversational platform allow for analytics and self-enhancements. Conversational AI technology ranges from the simple to the complex.

Path to conversational AI


Simple, information-focused chatbots are widely used and easily configured. The Dinner Ideas bot is a fun example of one. You type in an ingredient or meal, then it searches its database for relevant recipes and serves them up in Facebook Messenger. It’s pre-built, basic and only able to do spit out information culled from online recipes.


Transaction-focused virtual assistants are beginning to be adopted more and more across business verticals and functions. Capital One’s Eno, for example, helps customers by performing a number of specialized roles. It protects against fraud by alerting customers to suspicious account activity, while also providing financial guidance by keeping customers informed on their spending habits. Eno’s conversational AI has pre-trained intents and models that it works from to generate dynamic content for Capital One customers.


Context-focused virtual assistants normally built using a conversational platform reside on the more complex end of the conversational AI spectrum. As voice interfaces integrate with IoT, like BMW’s “Intelligent Personal Assistant,” the conversational AI-infused virtual agents can control our surrounding environment. Not only does BMW’s virtual assistant respond to free speech, meaning it’s continually learning from the interaction with the driver, it also controls the car, opening and closing windows, turning on high beams, etc. It’s a sophisticated blend of IoT, AI and voice interface.

Path to conversational AI

To achieve a more contactless customer experience—not to mention frictionless—businesses will need to think beyond chatbots, voice skills and smart speakers. The future is omnichannel, both multi-device and multi-modal, with highly contextualized and intelligent conversational AI-based virtual agents. Many enterprise companies create chatbots/voice agents in the absence of an overarching conversational AI strategy. Don’t fall into the trap of technology for technology’s sake.

Explore CX goals

An effective conversational AI journey begins and ends with the customer. By exploring avenues with specific business and consumer goals in mind, the technology will aim to please the people who matter most, your customers. While we’ve written extensively on CX elsewhere, what’s important to keep in mind is that in order to properly frame the solution, you need to first ask the right question, specifically who is your user and what experience do you want them to have.

Assemble cross-functional expertise

Involve a cross-functional team consisting of CX domain leaders from marketing, sales, customer service, operations and IT, and your chief data officer (CDO) to identify the business opportunities. The team should understand the business process and expected customer journey, have experience with conversational design, as well as how to integrate with existing IT systems for automation.

Create a vision

Prior to any development activities, initiate an exploration phase that focuses on how customers will use the virtual agents and what benefit they’ll derive from a chatbot/voice interface. While failing fast and iteratively improving is key to an effective ideation process, what you don’t want is to fail your customers with an end-product they have no use for. We facilitate virtual executive workshops to align to a future-state vision for driving end-to-end improvement of business outcomes. It’s these kinds of visioning exercises that work to identify innovative solutions and enablement requirements.

Start small and fail fast

Trying to do too much can result in a slower time to market. Rather than experiment across every possible audience and interaction channel, test bots with a single audience segment or interaction type. Then analyze, train and expand from there. Allow the conversational platform to grow organically by building on what’s working while shifting away from what’s not. Leverage existing data for intent, content mapping, including knowledge bases and historical data from existing client-facing apps, if available.

Continuous feedback

To deliver conversational AI that meets the evolving needs of your customer, you’ll need to iteratively improve on your approach. Ensure processes and practices are effective by gathering and analyzing metrics at regular intervals, which could include insights like user engagement, virtual agent resolution vs. escalation to human and cost of engagement while striving towards continuous improvement.

While the ongoing pandemic may have accelerated the contactless economy, the growing demand for Conversational AI will outlast the moment. The key is to be able to deliver the kind of technology that not only make consumers’ lives easier but also heightens the customer experience. With the market awash in low-end chatbots and voice agents that often undermine simple asks, consumers are beginning to expect more from Conversational AI. Just having a bot is no longer a differentiator. Companies need to develop Conversational AI with a clear path towards business value and the contactless, frictionless experiences consumers increasingly need.

Check out our predictions on the future of voice interfaces.

About the Author

Raja RoyRaja Roy is vice president of technology at Concentrix Catalyst. He is responsible for driving the overall technology initiative, strategy, and innovation for Concentrix Catalyst. Raja has nearly 20 years of experience leading technology strategy, initiatives and innovation for Fortune 500 brands.

Tags: , ,