Gartner Inc. predicts that organizations will develop 80% of Generative AI (GenAI) business applications on their existing data management platforms by 2028. This approach will reduce the complexity and time required to deliver these applications by 50%.
During the Gartner Data & Analytics Summit taking place in Mumbai this week, Prasad Pore, Sr Director Analyst at Gartner, said, “Building GenAI business applications today involves integrating large language models (LLMs) with an organization’s internal data and adopting rapidly evolving technologies like vector search, metadata management, prompt design and embedding. However, without a unified management approach, adopting these scattered technologies leads to longer delivery times and potential sunk costs for organizations.”
As organizations aim to develop GenAI-centric solutions, data management platforms must evolve to integrate new capabilities or services for GenAI development, ensuring AI readiness and successful implementation.
Enhancing GenAI Application Deployment With RAG
Retrieval-augmented generation (RAG) is becoming a cornerstone for deploying GenAI applications, providing implementation flexibility, enhanced explainability and composability with LLMs. By integrating data from both traditional and non-traditional sources as context, RAG enriches the LLM to support downstream GenAI systems.
Most LLMs are trained on publicly available data and are not highly effective on their own at solving specific business challenges,” said Pore. “However, when these LLMs are combined with business-owned datasets using the RAG architectural pattern, their accuracy is significantly enhanced. Semantics, particularly metadata, play a crucial role in this process. Data catalogs can help capture this semantic information, enriching knowledge bases and ensuring the right context and traceability for data used in RAG solutions.”
To effectively navigate the complexities of GenAI application deployment, enterprises should consider these key recommendations:
Evolve Data Management Platforms: Evaluate whether current data management platforms can be transformed into a RAG-as-a-service platform, replacing stand-alone document/data stores as the knowledge source for business GenAI applications.
Prioritize RAG Technologies: Evaluate and integrate RAG technologies such as vector search, graph and chunking, from existing data management solutions or their ecosystem partners when building GenAI applications. These options are more resilient to technological disruptions and compatible with organizational data.
Leverage Metadata for Protection: Enterprises should leverage not only technical metadata, but also operational metadata generated at runtime in data management platforms. This approach helps protect GenAI applications from malicious use, privacy issues and intellectual property leaks.
