By Balakrishna D R, Senior VP, Head – AI & Automation Services, Infosys
As the pandemic-induced normal continues to evolve, two aspects of business have changed fundamentally, and are reshaping the business landscape. One, the rewriting of the customer-business relationship in terms of what the customers expect from a brand (tangibles and intangibles) and how a brand responds. Two, accelerated adoption of mobile and digital technology among consumers as well as emerging technologies among enterprises, specifically artificial intelligence (AI).
This wave of technology adoption is blurring the line between customer experience and digital experience. And the finesse with which companies use AI to walk this line will be key to survival and long-term value creation. A point corroborated by a recent IDC survey which revealed that early adopters of AI reported a nearly 25 percent improvement on a number of parameters including customer experience, innovation, competitiveness, and business margins.
The untapped potential of AI for experience design
Owing to its industry-agnostic nature, AI has found applications across sectors for CRM. From traditional sectors such as telcos and retail to financial and health care services, and from e-commerce to other internet-enabled services. This has helped drive agent efficiency and operational efficiency for customer service functions.
And while this does contribute significantly to improving customer experience, a deeper application of AI to experience design in a data-driven world can help unlock incremental business value while delivering a more personalized seamless experience.
The many ways AI can serve the customer experience
As enterprises look to accelerate their digital transformation journey in a financially constrained landscape, decision-makers are open to exploring new solutions that go beyond the proof of concept, address some of the current challenges they are facing, and demonstrate business value. Here are some use cases for using AI to optimize design experience:
● Effective and efficient CRM: AI can help make CRM more efficient by using historical and contextual data related to incoming calls to route it to the relevant agent or issue-specific channel. Additionally, compared to the legacy systems, AI assistants can engage the customer for more effective self-service without having them go through complex processes or having to wait in queues.
● Intent prediction: Customer queries and conversations received by contact centers can be processed using machine learning (ML) and natural language processing (NLP) to perform sentiment analysis which can help better understand the customers’ intent and behavior. This provides businesses meaningful insights into the customers’ mind within a given context and can help improve the customer experience while enhancing the fulfillment capabilities of brands.
● Enhanced learning: Contact centers have a trove of voice and/or text-based conversations that can be analyzed and used to train agents. Using real-life calls, brands and CRM teams can develop multiple scenarios and simulations to improve the agent’s engagement with customers. Additionally, analysis of conversations with different customers can help create customer personas that can be used in training to humanize the exchange
Key considerations for successful AI implementation
The number of elements and variables involved in an AI environment makes implementation tricky. In fact, the earlier mentioned IDC study also estimates that 28 percent of AI/ML initiatives fail due to a combination of operational factors. Additionally, AI has also opened up a number of unique issues that organizations must focus on for successful implementation.
● Data management: Data generated from multiple sources comes in various shapes and formats, and as a basic input for AI, storing and preparing data for use requires massive infrastructure and investment to derive relevant actionable insights. Using a polycloud based approach makes managing the hybrid cloud infrastructure easy by providing a thin layer of abstraction over it to move applications, data, and enterprise workloads at scale – in a secure, consistent, and compliant manner.
● Biases: Technologies like AI and ML rely on historical data used as input and the algorithms applied for analysis. Even though AI models are designed to operate independently, some human intervention is unavoidable, and this is where bias comes into the picture. In order to avoid this trap, companies must invest in data cleaning and preparation. The key is to monitor the data preferably at the input stage, and definitely before it is processed because a biased dataset can skew the output due to extrapolation of ‘dirty’ data.
● Privacy: Numerous studies have revealed that the privacy of personal data in the hands of businesses is one of the biggest concerns among consumers. So, as we move into an era of personalization, businesses need to prioritize users’ data concerns and build in protection mechanisms within their digital ecosystem. Additionally, with privacy compliance regulations such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA), the onus now squarely lies on businesses to clearly define policies and governance mechanisms that ensure compliance in order to remain competitive.
The bottom line — despite all its potential, AI is still a relatively emerging field, and for an AI-based approach to fructify, enterprises need to deploy a multi-disciplinary approach where holistic customer experience and privacy concerns are a design mandate and not just an afterthought. This will enable brands to not only deliver personalized products and solutions that customers want, but also enhance the overall engagement and experience while remaining legally compliant and aware of privacy concerns.