Silicon Valley Innovation Center
We help global corporations grow by empowering them with new technologies, top experts and best startups
Get In Touch
Our Location

Innovation and Tech Resources Hub

Access the latest trends, best practices, educational materials, and support services designed to drive technological advancement and innovative thinking in your organisation.

In 2017, Facebook triumphantly announced that over 100,000 chatbots were available on the Facebook Messenger platform. Focusing on rudimentary, structured queries, these chatbots failed to deliver on the promise of intelligent Star Trek-type intelligent conversational bots. It seems the journey to true conversational AI had undergone a false start. Today, the quest for truly conversational AI is one that tackles deeper challenges than just understanding what the input is and predicting a possible answer. This associative approach to conversational AI is but the tip of the iceberg.

Kunal Contractor is global director at Avaamo, a company that is building the next generation of conversational AI applications for the enterprise. The company, working with some the of the largest companies in the world, is attempting to crack the conversational AI conundrum, one that will unlock the power of conversational AI to drive down costs and increase customer and employee engagement. We recently sat down with Kunal to discuss what the vision of conversational AI is for the enterprise and what challenges enterprises will need to surmount to achieve revolutionary digital transformation.

Customer-facing Conversational AI

Gartner predicts that by 2020 customers will manage 85% of their relationship with the enterprise without interacting with a human. This prediction is predicated on the rapid rise in conversational AI driven by advances in deep learning, big data and predictive analytics. However, Kunal explains, to truly deliver what can be called the promise of conversational AI, fundamentally new technology must be built to perform multi-turn conversations and execute judgment-intensive tasks, just like humans. What Kunal is referring to as multi-turn conversations are interactions injected with slang, insinuations, references to past conversations, colloquialisms and other language factors that current chatbots cannot handle.

However, when implemented correctly, conversational AI can quickly supplant human interactions. Consider how difficult it is to ask a fellow human a certain question but then ask Google to search the same question with no reservations. The belief that robots are not judgmental will be a huge factor that drives the successful adoption of conversational AI in the enterprise. Other factors that will potentially drive the uptake of conversational AI will be the instantaneous access to data AIs have resulting in instant answers to complex questions, the belief that AIs do not lie, as well as access to the customer’s entire history, not just of purchases but conversations as well. The potential for deep context and unprecedented customer engagement serves as an incentive for enterprises to pay greater attention to conversational AI.

Employee-facing Conversational AI

To demonstrate the potential conversational AI has to impact enterprise employees, Kunal offers two illustrations. In the first illustration, an investment bank employee with 20 years of experience needs to confirm a detail to a customer. He can either dig into the system to surface that information, which may take long or ask a junior associate for the information. To avoid embarrassment, the employee will opt to put the customer on hold and search for the information. In the second illustration, Kunal talks of a large enterprise organization that receives over 15,000 password reset requests from employees each month. It takes each employee around 20 minutes to get their password reset resulting in the loss of a staggering 5,000 work hours each month.

In both instances, it is possible to see the cost implications of the actions taken by employees to remedy the situation at hand. Functional conversational AI can remedy these situations. In the first instance, the employee seeking information can ask the conversational AI assistant for the information, both avoiding embarrassment and cutting the time it takes to respond to the customer. In the second instance, Kunal says that with conversational AI, it is possible to cut down the password reset time per employee to under 27 seconds, saving the organization close to 98% of the time employees previously spent resetting their passwords. It’s clear from this illustration why Juniper Research forecasts that chatbots and other conversational AI will be responsible for cost savings of over $8 billion per annum by 2022, up from $20 million in 2017.

Last-mile Automation

Conversational computing is a rapidly evolving technology, and it does promise to change the way that we work and how a lot of business would interact with their customers, with their employees, stakeholders, and with a massive capability of impacting both the customer experience and reducing cost at the same time, says Kunal. He calls this application last-mile automation. This last mile, however, represents a frontier that is both pregnant with potential yet fraught with challenges. As it stands, Natural Language Processing (NLP), what current conversational AIs use, is the easier part. What is more difficult to achieve is Natural Language Understanding (NLU), a state that will make artificial AIs able to respond to more complex multi-turn inputs.

Current AI technologies, while adept at probabilistic computing, falter when it comes to causal computing. For instance, a banking AI can understand a bank transfer command based on similar inputs but may not understand a question that requires causative reasoning. That is, if a customer asks, “How can I improve my credit rating?” Such a question must reach far beyond simply regurgitating standard credit rating answers to delve into the customer’s financial history, purchasing trends, earning trends, and other data that require more than just an X=Y, therefore, Y=X approach to answer in a meaningful manner.

Catching the Conversational Ai Wave

Organizations on the quest for digital transformation cannot afford to ignore conversational AI. Gartner predicts that by 2019, 20 percent of brands will abandon their mobile apps in favor of building services on top of other existing platforms like Facebook Messenger, WeChat and WhatsApp. Gartner also predicts that AI will be the foundational technology that drives the next wave of these enterprise applications. While in the past the enterprise technology conversation was framed around having a website and mobile apps, says Kunal, today, we are in the world of an AI-first strategy. He is quick to point out that from history, any and every early adopter to get technology always gets to the top. Enterprises will, therefore, do well to start the journey towards integrating conversational AI into both their customer facing and employee facing interaction infrastructure if they are to survive the 4th industrial age, or as Kunal puts it, Industry 4.0.

VIDEO: Interview With Kunal Contractor


Most Popular

Exclusively via mail

Learning to Innovate -
Intelligence Series

We specialize in delivering to you the unique knowledge and innovation insights of Silicon Valley!

Technology advances as you read -
Learn to Innovate today!

Let us help you