Prior to the LLM, the primary use cases for citizen data scientists were developing dashboards, performing data discovery steps in new data sources, and ad-hoc queries. Today, the needs of business teams and data scientists have expanded to include developing RAGs, embedding knowledge in SaaS LLMs, and leveraging AI agents. Data teams must have APIs for primary data sources and knowledge repositories that can be leveraged for these and future use cases.
Sisense CEO Ariel Katz said, “Integrating LLM knowledge with enterprise data can unlock predictive insights and enable real-time decision-making, transforming information workers into proactive decision-makers and catalysts for innovation. “Data teams must evolve from gatekeepers to enablers, abstracting complexity and providing data API services that make it easy for creators of all types, whether pro-code, low-code, or no-code, to embed analytics.”
APIs aren’t just for accessing data sources. When data teams create visualization components, ML models, RAGs, and AI agents, the first step to providing services should be ensuring a robust and easy-to-use API.
Michael Berthold, CEO of KNIME, said it is important to have guardrails in place for data quality and access before deploying models into production. “Enterprises recognize that models can make poor predictions or leak sensitive information. “Effective tools help oversee data flow, model usage, and add safeguards to reduce these risks.”
Building a data marketplace to simplify search
Data teams should consider citizen data scientists as one of their end-user personas, but other business users with less technical proficiency should also be able to search and access data sources. Using data catalogs and data dictionary creation is an important first step toward achieving broader data access. By building a data marketplace, companies can seize the opportunity to expand their self-service data and AI programs.
“Multiple layers of IT and governance bureaucracy slow down data access and make it difficult to accelerate new innovations, improve supply chain logistics, and deploy innovative AI applications,” said Moritz Plasnig, Chief Product Officer at Immuta. As AI adoption accelerates, the key is no longer the killer app. Data is the new app. “The data team has the power to enable anyone in the organization to become a data consumer by fostering an internal data marketplace that automates discovery and access while providing enterprise-grade governance and security.”
In industries that need to integrate multiple primary, high-volume data sources for many departmental use cases, data marketplaces can be an accelerating capability. Companies in manufacturing, construction, energy and other industries can use data catalogs and marketplaces to aggregate and streamline the use of real-time data sources for decision-making in marketing, field operations, supply chain, finance and other departments.
“Data teams are essential in industries like manufacturing where data is abundant but difficult to explore,” said Artem Krupenev, VP of Strategy at Augury. “Their role is not just to enable data to be operated, but to support everyone to become a data scientist by ensuring data accessibility, ease of use, and impact,” he emphasized.
Developing data products that foster collaboration
Marketplaces aren’t just about discovering, accessing, and integrating data sources. Data teams can now view advanced dashboards, ML models, LLM competencies, and AI agents as data products and manage them as product development initiatives. Each product has defined customer segments, value proposition, and strategic goals, which can be defined in a vision statement and managed through a product roadmap.
“The data product concept has evolved from a buzzword to a core element of modern data-driven organizations,” said Pete DeJoy, SVP at Astronomer. “The analogy with physical products and supply chains helps bridge the communication gap between technical and non-technical teams by clarifying the end-to-end data lifecycle.”
As more business teams become data-centric and the importance of AI as a business capability increases, the line separating data teams and business teams is becoming blurred. The future of work requires data teams to redefine their mission and deliver enhanced data governance, DataOps, marketplaces, and data products that serve more departments and use cases.
dl-itworldkorea@foundryco.com
Source: www.itworld.co.kr
