Never before has the business spotlight been on database managers and professionals as it is at this time. Every business leader now wants a data-driven organization, as they recognize this is the key to competitive differentiation. Data monetization is seen as a vast new revenue source. AI—both operational and generative—relies on effective data management and quality. The ongoing transition to digital business demands real-time delivery of insights to end users and applications.
Industry leaders were asked to describe the technology area where they see the most change and how it is shaping the way data is managed and delivered.
AI—IN ALL ITS FORMS
Of course, no discussion of game-gaming technologies is complete without mention of the widespread impact of AI in all its forms—operational, machine learning, and generative. Overall, it’s a unifying force, said Sunil Senan, global head of data, analytics and AI for Infosys, who referred to this new architecture as “systems of intelligence.” This architecture is “integrating data across all organizational functions such as marketing, sales, finance, and others.”
This, in turn, is resulting in “enterprise-wide pools of data that provide the foundation for advanced analytics and forecasting models,” Senan continued. “_is allows companies to generate predictions and insights with enhanced accuracy.”
The most recent variation of AI—generative AI—continues to stir excitement across the industry. “AI-powered applications have the opportunity to offer contextualized, hyper-personalized, and responsive user experiences,” said Chin-Heng Hong, VP of product management at Couchbase.
Potential issues: A potential roadblock to the success of AI-powered, adaptive applications is “complex data architectures, built from collections of disparate, purpose-built databases,” Hong stated. “It’s important to understand that complexity is AI’s enemy, as it’s imprecise to begin with,” he said. “Using AI within a complex, multi-database architecture does not lead to better AI results.”
Data complexity also complicates a user’s ability to construct large language model (LLM) prompts, Hong added. “Data concerning a user’s profile, preferences and behavior attitudes, location, real-time activity, and more must be up-to-date and accurate at the moment of user interactions. The problem arises when these variables are gathered from multiple databases, creating the opportunity for mixing unrelated or contradictory information that could allow LLMs to hallucinate.”
“To build a solid data foundation for optimal AI performance, companies need skilled personnel, robust processes, and dedicated resources in place,” Senan opined. “This ensures data is properly maintained, securely protected, and optimized to drive tangible business value. It is also important for business leaders to consider where their data is currently being stored, as the sheer volume of data companies possess can be a deterrent of modernization.”
“Data breaches and noncompliance with regulations pose a growing threat to both individuals’ privacy and companies’ financial well-being,” said Senan.
“One of the largest potential roadblocks to the proliferation of AI technology is inadequate data preparation and management,” Senan continued. “Challenges around data quality, such as incomplete, inaccurate, or biased data, can severely limit the performance and reliability of AI systems.”
Business benefits: The impact of AI promises to be nothing short of transformational, especially for data managers and professionals and their business counterparts. “By offering a panoramic view of organizational data, executives are better equipped for well-informed decision making and driving strategic advantages over competitors who are operating with fragmented or incomplete data views,” said Senan.
AI also serves as a “powerful partner in the innovation process. As this technology continues to evolve and integrate further into our lives, we can expect even more disruptive and transformative applications that will redefine how businesses operate and create value.”
Just about every industry will benefit from AI-powered data applications. “For example, retailers and ecommerce platforms can provide more engaging and personalized customer experiences, which can lead to increased sales and retention opportunities,” said Hong. “Banking applications can make actionable recommendations based on individuals’ financial health, and, in turn, better manage risk, and more often, identify customers that qualify for other products and services.”
KNOWLEDGE GRAPHS
The increased integration of unstructured data into enterprises has created a knowledge explosion, and knowledge graph-based AI may help users navigate this new environment. A knowledge graph “combines the problem-solving capabilities of AI with structured, machine-readable knowledge representation,” said Bryon Jacob, CTO and co-founder of data.world. “It improves data explainability by providing a clear, auditable trail of how the AI arrived at its conclusions. This is crucial in the current landscape, where governance is a significant challenge due to the ‘black box’ nature of generative AI. Knowledge graph-based AI addresses this issue by enabling governance on the factual information used in AI responses, thereby making the system’s decision-making process transparent and accountable.”
Potential issues: “Many people are not familiar with what knowledge graphs are or how they can be used in enterprise AI applications,” Jacob cautioned. “Knowledge graphs are still a new concept for many, and the idea of combining them with AI solutions is even less well-known.”
Business benefits: Knowledge graph-based AI “can help to solve the problem of the black box nature of many deep learning models,” Jacob said. “By basing AI decisions on structured data, knowledge graphs provide a level of transparency and explainability that is important for regulatory compliance and ethical considerations. This can help to build trust and make it easier to manage AI-driven systems. KG-AI also allows businesses to make better use of their data, turning static data stored in warehouses into dynamic assets that can be used in real-time.”