AI’s role in customer service
Artificial intelligence (AI) has a key role to play in orchestrating the customer journey. AI and machine learning (ML) have transformed the way vertical markets have come to interact with customers and changed the future of customer service forever.
With multichannel and omnichannel marketing now a part of everyday life, turning to AI to enable a seamless and unified brand experience is only one aspect of the technologies required to create this experience. AI is a tool, though, not a standalone solution, and cannot be used to compensate for poor omnichannel customer experience.
If, however, it is a supporting part of a good omnichannel strategy, then the power of AI can be unleashed.
“Giving agents and call centre service representatives a complete, cross-channel view of every customer interaction at all times is the nirvana of using AI and ML in brand management,” says Pranay Desai, Head of Enterprise Marketing at Freshworks. “Using AI correctly can provide invaluable insights into the changing needs and preferences of customers, creating scalable customer journeys and delivering consistent positive experiences.
“Using AI to close any service gaps and having greater intelligence and knowledge means being able to measure fluctuating customer purchasing patterns. This means brands can more precisely decide and define service strategies.
“It’s also important to note that an AI implementation is only as good or bad as the data flowing into it. Successful integration of various data sources is vital. Once you have the use cases defined, ensure that the processes and systems in place are capable of capturing all the necessary data needed to perform the task at hand. For example, if you want AI to bring real-time customer profiles in front of your agents, it needs to capture data from multiple systems such as marketing automation, CRM, billing and operations.
“All data should be secured using multi-level security measures, including virtual private cloud, role-based access (including two-factor authentication) and AES 256-bit encryption.”
Desai explains that AI applications can be structured to support human agents to deliver effective customer service in different ways. Agents can be grouped by advanced skill sets required to answer complex queries (level three) when AI can take care of simpler, level one and two queries.
On self-service portals, AI can also flag topics that need new answers and old answers that need to be updated for title or content, based on past engagement with the answers. This results in a higher performing knowledge base and higher deflection.
Agent-assist bots help surface real-time customer context along with recommendations on the next-best-action to be taken. This reduces the time taken to train new agents, gearing them up to make them customer-facing faster.
“When allocating a task to a specific skill, routing is a part of business process automation (BPA), which works based on pre-set rules,” continues Desai. “The language barrier is overcome as AI solutions automatically detect the language within incoming tickets to assign it to the right person.”
He stresses that it is important not to focus on only one side of the brand-customer interaction. While customer-facing AI, like chatbots, can enhance a customer experience, it is important to apply AI in a way that helps both agent and customer.
He says it is risky using AI for telephonic interactions and issue resolution because prompting callers with automated numerical menus, while common, is frustrating to customers. Also with voice technology capabilities often lacking when it comes to accents and lexicons, it can be difficult for a voice-based bot to understand spoken words and tone.
Flaws can lead to miscommunication, prompting frustration and undesirable outcomes. Desai suggests that organisations bring AI directly to agents and let it ‘listen in’ on conversations – no matter over which communication medium, determine the query topic and assist the agent to provide answers.
Applications exist that can suggest responses for the agent, including articles, documents, FAQs or Web pages previously created about the issue at hand, and the agent can filter or choose responses best fit for the customer. Because the agent sees the recommended responses before they go to the customer, they can adjust them as needed, avoiding miscommunication and negative outcomes, staying in full control of the conversation.
Optimising pricing is also possible using AI and ML because channel and brand preferences, purchase histories and price sensitivity can be established. This needs to be underlined by a sound and effective supply chain infrastructure that enables faster speed to market.
Order tracking across each channel combined with predictions of allocation and out-of-stock conditions will reduce operating risks and this, in turn, will provide feedback on where there are process inefficiencies.
“We see cloud-based management and pricing apps as being easier to use than ever before, thanks to AI and ML, and we're seeing this fuelling rapid innovation in the customer service arena into the future,” concludes Desai.
Download a guide to AI in customer service here.