Experts discuss how to deliver what customers want through advancements in AI

The integration of artificial intelligence (AI) and human agents is crucial for providing a seamless customer experience. Meena Vishnampet, Conversational AI Practice Lead at Google Cloud envisions “a future where AI and human agents work together to optimize resources and provide high-quality service.” The most complex issues can be handled by trained human agents, while simpler issues can be effectively resolved by AI agents.
Natural language processing (NLP) techniques are used to transform raw characters into sentences, parts of speech, entities, and actions, which are represented as vectors. Similarly, images are transformed into various visual elements, also expressed as vectors. However, these techniques can encode biases, racism and deception contained in the training data.

Social sentiment analysis has used natural language AI to interpret how influencers drive specific market results, but sometimes social media sentiment is null and void, and a human element is required, such as in the case of electric utility customers struggling to pay their bills.

Karl Hampson, CTO Data and AI at Kin + Carta emphasizes the importance of understanding the customer's most critical concerns to provide a satisfactory response through AI customer interventions. He provides an example of Google's "rapid responder" model during the pandemic, where an AI "box" containing the resolution to the top three queries streamlined the experience for customers.

Poor execution is the main reason for failed AI customer interventions, as issues falling outside the AI model's parameters require quick identification and a hand-off to a human agent. Meena Vishnampet suggests three successful AI initiatives for enhancing customer experience, including:

1. A 24/7 conversational AI for quick resolution of customer queries
2. Using AI to assist customer service agents in the background, and
3. Creating efficiencies with a future-proof solution that analyzes conversational drivers and identifies customer sentiment and frustration to drive training and coaching for agents

Ultimately, getting the fundamentals right is crucial for effective AI customer interventions.

The COVID-19 pandemic has presented significant challenges for AI models, and identifying solutions that are future-proof is crucial. Abdul Khaled believes that AI challenges can reflect the customer's challenges with the brand. At E.on, the majority of call-ins are from customers questioning the accuracy of their invoices compared to their historical usage charges, which have been much lower until now. To address this, Khaled suggests providing options such as an invisible AI running a report to compare bills with previous months, identify potential government program cost savings, and confirm the accuracy of the bill. In regulated industries, safety and governance are crucial factors in the use of AI, and it is critical to invest in AI ethics. Organizations must understand how personally identifiable information (PII) is used, especially in personalized marketing and analysis, and consider geographical regulations from a regulatory standpoint.

According to Meena Vishnampet, 2023 is "The Year of AI" due to the remarkable progress made in NLP and the increasing pace at which AI is developing. Large Language Models (LLMs), which are neural networks trained to execute specific tasks such as predicting the next word or sentence, are becoming increasingly sensitive to the contextual relationships between elements of language.

Advancements will first unfold in the enterprise, with productivity suites like Slack and Teams, and with organizations building their own solutions. Overall, the panelists believe that AI nirvana - a completely frictionless journey - is closer than we think.