How NewDay’s Generative AI Assistant is Transforming Customer Service Efficiency

Customer Service AI Impact

Customer service is often the first point of contact between a company and its customers, and for many businesses, getting it right can make all the difference. That’s why recent developments from UK-based financial services company NewDay are attracting attention—not just in the finance world, but across industries looking to improve how they support customers. NewDay has introduced a generative AI-powered assistant called “NewAssist,” designed to help customer service agents respond more quickly and accurately to customer inquiries. What makes this particularly interesting is that it’s not just a concept or pilot—it’s already showing results, with over 90% accuracy in understanding and answering questions.

Generative AI Development

The idea for NewAssist began at an internal hackathon in early 2024. The challenge was clear: how could generative AI be used to help agents find answers faster from a large set of internal documents? With nearly 200 knowledge articles and over 2.5 million calls handled annually, even experienced agents sometimes struggle to locate the right information quickly. The team behind NewAssist focused on building a real-time assistant that listens to conversations using speech-to-text technology and then suggests relevant answers during live interactions.

Overcoming Initial Hurdles

Initially, the team faced several hurdles—limited infrastructure, competing priorities, and the need to prove that AI could actually add value. Rather than trying to build a full voice assistant from the start, they took a more manageable approach by creating a chatbot prototype. This allowed them to test their ideas in smaller steps while focusing on improving accuracy. Using Amazon Web Services (AWS) tools like Bedrock and OpenSearch Serverless, they created a system that retrieves relevant content from their knowledge base and uses a large language model (LLM) to generate helpful suggestions.

Data Quality’s Role

One key factor in their success was data quality. At first, they used general-purpose tools to process documents but found that results were inconsistent. By developing custom tools tailored specifically for their internal document format, they significantly improved how well the AI understood context—raising accuracy from around 60% to over 80%. Another important learning came when they started testing with real users. Agents often used abbreviations or informal terms when asking questions—something the development team hadn’t anticipated. By adjusting how these inputs were handled, accuracy improved again.

Broader Trends in AI

Looking at this in context, NewDay’s journey reflects broader trends we’ve seen in enterprise AI adoption over the past few years. Many companies have experimented with generative AI since tools like ChatGPT became widely known in late 2022. However, moving from experimentation to real-world application has been challenging for most organizations due to concerns about cost, reliability, and integration with existing systems. What sets NewDay apart is their careful step-by-step approach: starting small, focusing on measurable improvements, and adapting based on user feedback.

AWS Support for Generative AI

This also aligns with AWS’s broader strategy of supporting generative AI through scalable cloud infrastructure. In recent years, AWS has launched services like Amazon Bedrock specifically aimed at helping companies build custom AI applications without needing deep technical expertise in machine learning models themselves. By leveraging these tools effectively, NewDay has shown what’s possible when business goals are clearly defined and teams are empowered to experiment.

Conclusion on Customer Experience

In conclusion, NewAssist is more than just another AI tool—it’s an example of how thoughtful design and gradual implementation can lead to meaningful improvements in customer service operations. While it may not grab headlines like some of the more dramatic claims around AI replacing jobs or transforming entire industries overnight, its impact is real: faster responses for customers and better support for agents doing complex work under pressure.

Takeaways for Customer Roles

For readers working in customer-facing roles or managing support teams here in Japan or elsewhere, this story offers some useful takeaways. Generative AI doesn’t have to be overwhelming or expensive if approached carefully—and understanding your own data and workflows is just as important as choosing the right technology platform. As more companies follow similar paths toward practical AI use cases, we’ll likely see continued improvements not only in efficiency but also in overall customer experience.