AI will augment but not replace human judgement
March 26, 2018
According to Gartner, 55% of mature organizations have begun investing in AI or plan to do so by 2020. AI has moved from the realm of imagination into the world of reality. Chatbots have caught everyone’s fancy. Yet there is fear that AI may soon take over human jobs.
We met with Asheesh Mehra, Co-founder and CEO, AntWorks, to get his perspective on the subject. Antworks is a global AI company that works with automation and enterprise intelligence. AntWorks is the only provider of integrated intelligent automation enterprise-level product that is powered by Fractal Science.
/ / What are the current trends in customer service affected by automation?
The overarching trend in customer support has been self-service. Customers do not like to talk to customer service agents. Chatbots have had a long history of evolution. Gen 1 chatbots were hard-coded and dumb bots that provided predefined responses. Then came the smarter Gen 2 bots which depended on NLP to answer queries. We, at AntWorks, have been collaborating with several of our clients to build self-service bots that depend on NLP and even NLG (natural language generation) to fetch the required information and handshake with the enterprise system to fulfill transactional requests. AI-powered chatbots are getting smarter and smarter every day. Chatbots today can handle both routine and complex queries, so everyone wins. Customers get the answers they want and companies save costs along the way. AI will be leveraged to automate administration and to augment, but not replace, human judgment.
/ / What strategies should businesses adopt to transform outdated customer service touchpoints into omni-channel, AI-enhanced processes?
AI is ideal for omnichannel conversations. Consumers expect conversations they have with an enterprise to flow from one channel to another, so they don’t have to backtrack or repeat themselves. An AI-powered virtual assistant could be used very effectively in this context. The assistant could also feed human agents relevant facts to provide the right answers.
/ / How is AI helping to augment the work of human customer service representatives?
AI will be leveraged to automate administration and to augment, but not replace, human judgment. AI-powered assistants will anticipate needs by context, preferences, and prior queries and will deliver proactive alerts, relevant offers, or content. They will additionally become smarter over time due to artificial intelligence.
/ / According to Forrester, AI will replace 7% of US jobs by 2025, leading to a fear that it will render some careers obsolete.
How can organizations help their workforce adapt to these changes? While we call it Artificial Intelligence, it’s actually Acquired Intelligence. This intelligence in the machines is acquired from a human. This teamwork between humans and machines will be the most powerful combination in the world. As I said, AI will be leveraged to automate administration and to augment human judgement. However, AI will bring with it new criteria for success: collaboration capabilities, information sharing, experimentation, learning and decision-making effectiveness. Organizations will have to develop training and recruitment strategies for creativity, collaboration, empathy, and judgment skills. Enterprises will have to develop a diverse workforce and team of managers that balance experience with creative and social intelligence — each side complementing the other to support sound collective judgment. Setting new KPIs can be leveraged to drive adoption of AI.
/ / Google has predicted that its AI-driven assistant will become a “predictive, all-knowing, super helpful and conversational assistant.” Are there any limitations with AI?
This question reminds me of Microsoft shutting down a bot called Tay after pranksters pushed it to make racist, sexist and pornographic remarks. Machine learning works best in an environment with rules and huge numbers of data points. The minute things get fuzzy—either due to a lack of rules, an unclear evaluation of success or a lack of data—artificial intelligence performs poorly.