Data accessibility, though crucial for maximizing data’s value, comes with a laundry list of requirements. From striking the balance between openness and compliance to establishing proper access controls, data trust and data lineage, and ensuring high scalability and availability, data democratization is both an essential and daunting task.
All hope is not lost, however; DBTA’s latest webinar, Enabling Self-Service for Data Democratization, gathered experts to explore the technologies and best practices shaping how enterprises achieve true data democratization amid its steep requirements.
In the last 20 years, “data has been transforming drastically,” as Catalina Herrera, field CDO, Dataiku, pointed out.
Data, once exclusively trapped in silos and managed by specialized teams, no longer fits this highly rigid structure. As the eras of Big Data and AI have altered the way enterprises interact with their data repositories, one thing has become clear, according to Herrera: “Self-service is not just a convenience, but a competitive advantage.”
Yet there are still challenges impeding self-service data democratization. The skills gaps between technical and business teams, rigid legacy infrastructures, and data access and quality issues, Herrera explained, are the main causes of friction.
Successful self-service data democratization is a series of moving parts, where diverse infrastructures reign. At its center lies a secure, governed, and scalable framework from which insights are derived from a range of data sources. Crucially, on top of this framework, business users and the applications they rely on must be able to access this information in a way that feels easy and intuitive to them.
Additionally, while democratization is the goal, control is key to its success. Understanding data lineage and all the ways it is interacted with balances openness with compliance while unifying teams and empowering their collaboration.
Of course, “the only constant is change,” said Herrera, and AI is aggravating that evolution exponentially. Data science will not be the same in five years, and AI-powered decisioning systems are already shifting focus. While currently, human analysts provide decision support to lines of business manually, eventually, AI-powered decision systems will be built to provide instantaneous, always-available insights—and large language models (LLMs) are at the center.
LLM-powered applications are different from traditional applications in both their design and capabilities, and enterprises will need to design, build, and maintain many LLM-powered applications to get ahead of its competition, Herrera predicted. Natural language interaction with data is becoming more and more expected, and preparing for that reality now will charge enterprises' abilities to deliver later.
“An educated consumer is your best customer,” said Danny Sandwell, technology strategist, erwin by Quest, emphasizing the role that data democratization has on positive business outcomes. Data democratization with control, paired with context, concise descriptions, a view into data quality and its applicability to specific use cases is the standard. But how do you get there?
Sandwell introduced a customer success story at St. James’s Place, a British financial advice and wealth management company, that employed erwin Data Intelligence by Quest to tackle the auditability of their data.
While adoption of erwin by Quest’s technology achieved St. James’s Place’s original goal, they had realized something else: “erwin Data Intelligence is more than a reference library. It's a living, breathing capability that will allow us to manage our data estate much more effectively in the future,” said Ian Peters, divisional director of group data management at St. James’s Place.
erwin by Quest enabled St. James’s Place to establish a single source of truth around its data repository, fueling its AI journey. Through the following framework, erwin by Quest enables the democratization of trusted, AI-ready data:
- Model: Design data architecture.
- Catalog: Search and find data easily.
- Curate: Enrich with business context.
- Govern: Apply business rules and policies.
- Observe: Raise data visibility and integrate data quality.
- Score: Automate data value scoring.
- Shop: Make trusted, governed data widely accessible.
Each of these steps help cultivate a system of observable, actionable data intelligence. After all, “if you want to get intelligence from your data, you need to be able to get intelligence about your data,” said Sandwell. Data models, data cataloging, data lineage, semantic views and knowledge graphs, impact analysis of data changes, and both establishing and expressing data quality are all at the forefront of a successful data democratization journey.
This is only a snippet of the full Enabling Self-Service for Data Democratization webinar. For the full webinar, featuring more detailed explanations, a Q&A, and more, you can view an archived version of the webinar here.