The Challenge

Efficiently managing and querying vast datasets to enhance customer support.

Yes Energy needed a solution to efficiently manage and query their extensive collection of datasets to provide timely and accurate responses to customer inquiries. The company faced significant challenges in locating the right data types containing the necessary information, which hindered their ability to support energy sector clients. To address these challenges, they needed to establish a robust data querying system that would:

• Quickly ingest, structure, and query large volumes of data.

• Provide timely and accurate responses to specific customer needs.

• Enhance the overall efficiency and responsiveness of their customer support.

The keys to this solution involved leveraging advanced AI capabilities to streamline data querying processes and tackle challenges the company historically faced, which included:

• Difficulty in locating relevant data types.

• Inefficiencies in data retrieval processes.

• Inconsistencies in data management.

• By addressing these challenges, Yes Energy aimed to significantly improve their customer support operations and better serve their energy sector clients.

The Solution

Developing an AI-powered API for efficient data querying.

To address the challenge of efficiently managing and querying their extensive datasets, Yes Energy partnered with Credera to develop a robust API leveraging generative AI and natural language processing. The project began with a comprehensive analysis of Yes Energy's data landscape and support needs. Credera's team refined the problem statement, conducted user research, and identified key requirements to ensure the solution would effectively meet the needs of Yes Energy's support team.

The solution involved ingesting and organizing 1.4 million records into an Opensearch Serverless vector store, enabling efficient data retrieval. Credera then developed an API that utilizes Amazon Bedrock generative AI to query these records and generate augmented responses. The development phase included creating prototypes and iterating on the models and retrieval augmented generation (RAG) output to ensure usability and functionality.

To ensure the API's effectiveness, Credera implemented an evaluator to continuously test and improve the accuracy of the AI-generated responses. Additionally, the team focused on seamlessly integrating the API with Yes Energy's existing systems to ensure minimal disruption and maximum benefit.

Following a product-focused approach, Credera ensured the project was user-centric, aligned with business goals, and collaborative. The team's expertise in product management facilitated the prioritization of key features and alignment with Yes Energy's objectives. This approach allowed for quick adaptation to new requirements, effective communication, and collaboration across cross-functional teams, as well as the identification and mitigation of risks to ensure continuous and timely feature delivery.

The Results

Transforming customer support with generative AI.

Previously, finding the right data involved time-consuming processes and reliance on tribal knowledge. Now the AI-powered API has streamlined this into a simple, efficient process of querying and receiving instant, accurate results. This transformation has positioned Yes Energy to better meet the needs of their clients, enabling them to provide timely and accurate information, resolving many customer requests immediately that previously required a team member with tribal knowledge to investigate.

The generative AI capabilities have enabled the system to deliver relevant and augmented responses to a wide range of customer inquiries, significantly enhancing the overall customer experience. As a result, the solution accurately responded to 80% of customer queries, showcasing the transformative potential of AI-driven insights.

Moreover, leveraging Credera’s Data Marketplace accelerator, Yes Energy successfully ingested and structured 1.4 million records within a short proof-of-concept timeframe. This rapid deployment has demonstrated the potential of generative AI to revolutionize customer support and operational efficiency within the energy sector.