Say Goodbye to Knowledge Base Hassles: Meet the CSR Chatbot Solution

AI themes inhabited the room at nearly every customer service and engagement workshop at E Source’s Forum 2023 in Denver this year. Speakers, practitioners and leaders expressed thoughts and questions about the value and use of Generative AI like ChatGPT in their organizations and touching their customers. All of the utilities present were considering AI in some capacity, but none of the utilities we spoke to at the event wanted to expose LLM powered chatbots directly to their questions. Everyone seemed more comfortable in a “wait and see” perspective related to Generative AI.

We’ve written frequently about the advantages that tools like ChatGPT and Generative AI poses for the utility industry. Given the natural skepticism or concerns about implementing a customer facing Chatbot, we advocate utilities consider an economical and effective alternative: a chatbot for Customer Service Representatives (CSR) in the support center.

Instead of building a customer facing chatbot first, leverage the inherent capabilities of Generative AI bots to replace a knowledge base for CSRs. A CSR chatbot requires all of the same steps as a customer facing application, with the added security and protection of an expert QA team to test and evaluate the effectiveness of the chat interactions months before any direct pilot with customers. Utilities can learn the process of a LLM chatbot implementation including data integration, training and priming, UI/UX customization, performance evaluation, data protection and security, and ongoing support for an internal team of expert CSRs that provide an additional and effective layer of control and feedback that utilities clearly require to build confidence in the technology.

A CSR Chatbot solution would replace existing knowledge base tools and overcome the inherent limitations of that technology. Most contact centers and customer support teams spend too much time, money and resources on maintaining their CSR knowledge base only to struggle with persistent challenges:

• Content Accuracy and Relevance: It takes too long for new information to permeate the knowledgebase and get into the hands of the CSR teams, making a knowledge base almost instantly outdated despite constant grooming and updating from a content team.

  • Generative AI learns from every interaction with constant feedback loops and can be trained on new information as it becomes available, which means the next chat question is already informed by new information.

• Content Organization and Searchability: legacy knowledge base applications require a librarian’s mentality for content organization and searchability and most CSRs have to learn the artificial taxonomy of the content hierarchy to find even the most common information.

  • LLMs thrive on content searchability by making normal language connections between facts, just like how customers “think” about the material, which maximizes searchability for information. Chatbots make multiple connections to the same information so that CSRs can pose different types of questions to get the correct and accurate information to give to customers.

• Training and Adoption: CSR teams typically devote a lot of time to learning HOW to use a knowledge base by learning the hierarchy of categories and content and how to search and find the information their customers require on a call, adding expense and new skills to training.

  • Generative AI chatbots use plain language questions to process questions, not a synthetic formal query, making it far easier for agents to work with the CSR Chatbot intuitively and without much training, and speeding user adoption to these powerful tools.

• Multilingual Support: Building a knowledge base in English poses inherent challenges and obstacles, so translating it into all of the languages our agents and customers speak natively adds layers of complexity and cost.

  • Most Generative AI chatbots can compose multilingual responses in real time depending on the inquiry through a process of priming and training the AI for new languages, which means the CSR chatbot can support multiple languages faster and more affordably than a traditional knowledge base.

A CSR Chatbot offers Customer Service Representatives with a real time support system with plain language capabilities that improve the accuracy and consistency of the customer service response along with CSR efficiency. Chatbots provide CSRs with an instantly available answer to customer questions that enables plain language inquiries with fast and accurate responses to any customer issue.

Training the CSR Chatbot with the essential information Customer Service teams require begins with an update of the current knowledge base to give the chatbot a base of information. Layer in new policies, processes and program documentation to keep the CSR Chatbot current and learning on an ongoing basis. Let an experienced team of CSRs test the chatbot with recorded customer calls to measure CX and data integrity, and establish performance benchmarks, then pilot that group of agents on live calls with customers. With the advantages of Chatbot plain language search, ongoing training and user adoption, CSRs will likely transition swiftly to these powerful applications.

Utilities already face demanding customer expectations for the digital and customer service experience based on the CX innovations of companies like Amazon, Apple, and Google. Companies in banking, healthcare, retail and ecommerce have already started incorporating Generative AI and advanced chatbots into their customer service experiences. Since utilities share customers with all of those brands, the industry will surely see customers expecting the capabilities of a chatbot soon.

So, Generative AI is coming whether we’re ready or not, and it makes sense for utilities to find a path to test and implement this technology that conforms to the standards of safety, security and control that our institutions demand. We encourage utilities to consider a CSR Chatbot as a means of learning, testing, and including this technology in a customer service capacity that innovates for the CX, but with the control and security the utility and its regulators require. For more information about Chatbot tools and strategies for including Generative AI in your organization, please contact us.