OpenAI’s ChatGPT, Google’s Bard, and other generative artificial intelligence (AI) systems represent a significant development for customer service and customer experience (CX) management due to their advanced natural language processing (NLP) capabilities, which can produce automated and detailed responses to customer service requests.
Generative AI systems employ large language models (LLMs) based on deep learning neural nets that have been pretrained on large corpora of data. These large, foundation models — consisting of hundreds of billions of parameters — have led to increased accuracy in NLP. Their general availability allows developers to build systems that can perform significantly more sophisticated NLP tasks than previously possible.
For example, when it comes to natural language understanding and natural language generation, LLMs can understand the subtle nuances in users’ conversations and generate much more coherent — and syntactically and grammatically correct — document-length text output. Moreover, LLMs like GPT-3 — the model powering ChatGPT — can “remember” the context of a conversation — even when it is conducted in an ongoing, iterative fashion. This is significant because (unlike earlier NLP systems) it means that their output does not become repetitive or redundant, even during lengthy or repeated user interactions regarding the same topic of conversation.
Integrating Generative AI with Customer Service Platforms
Integrating generative AI systems with customer service platforms will enable service-automation systems to provide much more extensive and detailed responses to customers’ questions than is possible with the current crop of such systems operating on their own. This includes providing automated (i.e., self-service) advice to customers and advice to help support human customer service reps (CSRs) perform their jobs.
We are already seeing this. Specifically, customer service automation vendors are now enhancing their platforms to allow generative AI systems to access the knowledge bases of their customer service platforms and ticketing systems. Such integration is key because it allows systems like ChatGPT to automatically respond to customer service tickets by analyzing the ticket’s content — using intelligent search, summarization, and language-generation capabilities — based on its training data and company-specific information. Benefits include reduced customer wait times and an improved CX and better customer service in general.
For example, customer service automation vendor Y Meadows has integrated ChatGPT with its platform. This means that customer-support teams can now directly connect ChatGPT with their internal knowledge bases and ticketing systems — including those in Salesforce and Zendesk. ChatGPT integration also supports direct connection with Microsoft Outlook and Google Gmail. This integration allows CSRs to immediately respond to customer emails and customer-support cases using ChatGPT-generated content.
Functionality-wise, the Y Meadows platform matches the customer’s question with the relevant knowledge base, including a company-specific knowledge base. Then, ChatGPT will generate a response to the customer’s issue and integrate those responses with the appropriate information from the company’s internal support system. For applications requiring human approval, Y Meadows supports “human-in-the-loop” verification of a response.
Generative AI and Virtual Assistants & Chatbots
Organizations can integrate generative AI systems with virtual assistants and chatbots to provide automated, detailed responses to customers’ questions or to guide them through processes designed to allow customers to solve problems on their own. Such functionality can improve the speed and efficiency of customer service, offer 24/7 customer support, and free up human CSRs to focus on more pressing customer needs.
Organizations can also use virtual assistants integrated with, for example, ChatGPT to provide personalized advice and recommendations to customers based on their service and purchasing history. This can help improve overall customer engagement and the CX.
Home warranty and home services provider Puls is developing a generative AI assistant to understand and respond to customers’ home-care questions. The assistant will provide real-time customer service support — including personalized guidance as well as pricing estimates for home warranties and work-time estimates for home-service calls.
In addition to providing instructions and advice to customers, the assistant will share information with the appropriate service technician. This includes giving technicians an accurate description of the service call request (e.g., type of repair), which is expected to reduce the need for multiple service visits and help avoid long wait times for parts.
AI Code Generation for Ad Hoc Data Access
Some vendors are using generative AI technology to create natural language interfaces that can generate SQL and other code from user-written prompts to access company databases and other knowledge stores. Companies can use these natural language text-to-code generators to support employees as well as customers — including with self-service applications.
Seek AI has developed a natural language interface that can generate code from users’ questions to instantly query databases. The interface allows business users to ask ad hoc questions and quickly get answers. It also helps reduce the burden of repetitive coding demands on IT/data access teams. Users can access the interface via email, Slack, text, and customer relationship management systems.
Seek AI’s interface seems particularly well suited for use by customer success teams that need to answer large numbers of varying customer questions on a daily or regular basis.
Organizations should investigate the use of generative AI systems for customer service and CX in general. The bottom line: these systems are expected to significantly increase the ability of customer service platforms — including those for automating customer self-service — to provide detailed responses to customer service requests. Generative AI systems are also widely applicable for supporting and enhancing customer engagement as well.
The industry is moving rapidly to meet organizations’ needs: already we are seeing vendors announcing new customer service products employing LLMs. And such developments are expected to ramp up considerably over the next few years, with vendors increasingly integrating ChatGPT and other generative AI systems with chatbots, virtual assistants, ticketing, and other customer service/customer engagement systems, and CX platforms.
Finally, I’d like your opinion on the use of generative AI systems in customer service, customer engagement, and CX. This includes any issues (e.g., intellectual property, privacy, security) you see arising from enterprise use of the technology, As always, your comments will be held in strict confidence. You can email me at firstname.lastname@example.org or call +1 510 356 7299 with your comments.