The #1 question to reduce AI costs
High AI costs are often driven by fragmented data that forces models to expend excessive tokens trying to interpret missing business context. Implementing a semantic layer provides…
Back to all articles
Category
Data Operations
20 articles · Page 1 of 2
The 90-second data model: Why AI-generated data models aren’t a modeling practice
While general-purpose AI can quickly generate plausible data models, these isolated snapshots lack the governance, connectivity to live databases, and security required for enterpr…
Replacing PowerDesigner: Real migration & modernization success stories
As SAP PowerDesigner approaches its end of life, many leading organizations in highly regulated sectors like finance and aviation are migrating to Quest Software’s erwin Data Model…
25 customer meetings in 3 days: 5 recurring themes about data and AI in 2026
After analyzing 25 customer meetings across various industries, Glenda O’Keefe identifies five recurring themes regarding the challenges of scaling AI and data products. While tech…
AI won’t fix a broken data foundation: It amplifies it
Data modeling is the essential foundation for successful AI implementation, providing the shared semantic definitions required for effective data and AI governance. Without this fo…
Lost in translation: Why semantic chaos is the real AI risk
The article warns that "semantic chaos"—the lack of shared business definitions—poses a major risk to enterprise AI by enabling models to confidently scale errors and drive costly …
Data product reusability: Cut costs & scale AI with 3 key changes
Many organizations waste significant time and resources by rebuilding existing data assets due to poor discoverability and a lack of trust in current datasets. To break this cycle,…
The cost of standing still: what data inaction really risks in an AI-driven market
Resisting change in data management leads to lost competitive advantage because fragmented, ungoverned data prevents organizations from reacting quickly to market shifts. To remain…
The trust paradox: Why better data programs are still rebuilding from scratch
Recent research reveals that even organizations with mature data programs are rebuilding assets instead of reusing them due to a fundamental lack of trust in existing data quality …
AI doesn’t have a data problem. It has a navigation problem.
The article argues that enterprise AI success depends less on acquiring more data or larger models and more on implementing a "navigation layer" of business context to ensure data …
From data access to data empowerment: why AI changes the data product equation
Many organizations struggle to derive value from data because their existing datasets are not designed for practical use, leading to inefficient, manual preparation processes. To b…
Snowflake heist: How to crack the code on Snowflake ROI
To maximize Snowflake ROI, organizations must prioritize proactive architectural design and preparation rather than relying on reactive cost-cutting. By implementing precise data m…