"A lakehouse combines the scalability and flexibility of a data lake with the governance, structure, and performance of a data warehouse. It allows organizations to store both structured and unstructured data in one platform while supporting robust analytics, machine learning, and BI workloads. These lakehouses get by the data swamp problem by providing features that are database like. That includes support for ACID transactions." (Fern Halper, "Data Makes the World Go 'Round", 2026)
"AI maturity depends as much on people as it does on tools. Organizations need to build AI literacy across roles, from analysts to executives, so that teams can ask the right questions, interpret AI results, and act on them. Cross-functional teams with skills in data engineering, machine learning, business domain knowledge, and governance are essential. AI systems should be treated as products and not one-off projects." (Fern Halper, "Data Makes the World Go 'Round", 2026)
"Data products are part of the enterprise data infrastructure. They are managed, versioned, and monitored much like APIs or microservices. This shift is supported by modern cloud-native infrastructure, which enables decentralized development, consistent governance, and scalable consumption. Data products don’t need to be stored in one location. They might be in a data lake, a data warehouse, a database, or a data lakehouse. They can become discoverable and accessible through a centralized data catalog or marketplace acting as a portal where users can find, understand, and request access to relevant data products." (Fern Halper, "Data Makes the World Go 'Round", 2026)
"Data readiness also involves the quality, integrity, and findability of the data. Is the data trustworthy and curated? Are data pipelines (to move data from source to target) and access controls (to ensure the right people get access to the right data) in place? Moreover, to be ready, the organization’s architectural components - data lakes, warehouses, lakehouses, etc. - must be coherent and aligned to support modern AI applications." (Fern Halper, "Data Makes the World Go 'Round", 2026)
"If AI is owned solely by IT or the data science team, it will fail. AI must be co-developed with business stakeholders. This means engaging stakeholders from the start, co-owning key performance indicators (KPIs), and embedding AI into workflows, not bolting it on after the fact. AI teams need domain experts, business sponsors, and clear lines of feedback. Use cases should be selected with the business, not handed down from a tech function." (Fern Halper, "Data Makes the World Go 'Round", 2026)
"One of the most overlooked aspects of AI deployment is the operational side, what happens after a model is developed. Operational readiness is concerned with the systems, processes, and team structures required to bring AI into production. In many cases, organizations build prototypes that never make it past the lab. To succeed, organizations must have formalized processes for deploying models, integrating them into business workflows, and monitoring their performance over time (no model is good forever; they degrade as the external environment changes)." (Fern Halper, "Data Makes the World Go 'Round", 2026)
"Organizations should use a framework (such as business value vs. complexity) to identify high-impact, feasible AI projects. Organizations that succeed with AI don’t chase hype. Instead, they score and prioritize use cases based on criteria such as business value, data readiness, effort, and complexity. This helps avoid wasted time on technically interesting but low-value projects. Some create the frameworks themselves to determine the high-impact yet feasible projects. Others rely on consultants to help them with this." (Fern Halper, "Data Makes the World Go 'Round", 2026)
"The modern data environment includes structured data as well as text, machine-generated logs and events, images, and audio. The rapid adoption of generative AI (GenAI) has accelerated both the creation and consumption of text data, increasing volume and variability. Additionally, hybrid and multicloud architectures are now standard. Data are stored and processed across warehouses, lakehouses, SaaS platforms, and edge services. Governance must operate consistently across these environments. In practice, this requires policies that can be enforced programmatically, with end-to-end lineage (where the data came from and how it has been changed) that is visible to both technical and business stakeholders." (Fern Halper, "Data Makes the World Go 'Round", 2026)
"[...] organizations are working to unify the data silos that often exist across their data ecosystem. Some are doing this through a centralized physical approach, such as implementing a data lakehouse pattern, which allows them to store and manage all types of data in one place. Many lakehouse implementations such as those from vendors such as Snowflake and Databricks have evolved into what is often referred to as the modern data platform. This is an architectural pattern that combines the lakehouse with tightly integrated tools for data ingestion, transformation, analytics, observability, and governance. This pattern reflects a shift toward cloud-native architectures that are designed to support end-to-end data workflows with scalability and flexibility." (Fern Halper, "Data Makes the World Go 'Round", 2026)
No comments:
Post a Comment