Practical guidance on data governance, AI governance, software architecture, and the future of agentic AI development. Written by Joshua Garza.
While the industry watches the model race, one argument says the lasting advantage sits in the runtime underneath: the layer agents must pass through to act on a device. Here is what that would mean for builders, including those handling regulated data.
Your best AI prompts vanish every session. Agent skills make them permanent — here's how knowledge workers are closing the productivity gap.
Dark factories run without humans on the floor. Agentic AI is doing the same thing to software operations — and the governance has to be built before the humans step back.
The EU moved the AI Act's high-risk deadlines to December 2, 2027 and August 2, 2028. Here is what US companies need to do now.
You cannot improve what you have not measured. Here is what a data governance maturity assessment looks like.
Vague complaints about technical debt rarely unlock budget. This guide shows engineering leaders how to measure debt systematically, translate it into business metrics, and build a paydown roadmap that leadership will fund.
Unit tests cannot capture emergent behavior in agentic AI. Learn how scenario-based validation gives teams the reliability signal they actually need.
Before an autonomous agent touches production, it should prove itself in a faithful replica. Here is how to build one and graduate agents safely.
Block connected its open-source Goose agent to DataHub's MCP server to give engineers metadata context during incidents, discovery and impact analysis. Here is what that pattern means for organizations with distributed data.
Ad-hoc data quality checks collapse under their own weight as data volumes grow. Learn how to design a rules engine with structured rule anatomy, five common quality dimensions, execution patterns, a maintainable rules catalog, and scorecards that track quality over time.
Learn how to implement data observability across the five core pillars, define meaningful data SLAs with business stakeholders, and build an alerting and incident response workflow that keeps pipelines trustworthy.
Data catalog initiatives often stall not because of technology, but because of misaligned strategy. Here is how to build one that people actually use.
Many organizations collect data for years before anyone asks what they're doing with it. Here's how to build a practical data strategy from zero — one that executives will fund and teams will actually follow.
Data governance initiatives often stall not because of bad tooling, but because no one owns the data. Here's how to build a stewardship program that creates accountability, improves quality, and actually sticks.
Bias testing is becoming a legal duty for high-risk AI. Learn the three types of AI bias, proven fairness metrics, when to audit, and what documentation regulators expect.
Direct database access is not an API strategy. Here is how to design deliberate, durable API boundaries for data platforms — covering protocol trade-offs, versioning, pagination, security, and contract-first development with OpenAPI.
You cannot govern what you have not inventoried. Learn how to build a defensible AI system inventory, classify risk tiers aligned with the EU AI Act, and maintain the registry as a living governance document.
Maturity assessments, policy development, stewardship, data quality culture
EU AI Act, NIST AI RMF, AI system inventory, bias testing, compliance
Technical debt, modernization, microservices, API design, cloud migration
Dark factory principles, scenario-based validation, autonomous agents, Digital Twin environments
Data quality rules, observability, monitoring, SLAs, tools and frameworks
Data strategy, technology leadership, team building
Subscribe to The Touchstone Brief -- a monthly newsletter with practical governance and architecture insights. No spam, no fluff.
Newsletter launching soon. In the meantime, follow us on LinkedIn for regular insights.