AI and Data
A starting point for thinking about where generative AI can help data professionals, and where messy data still needs human judgement.
Category
AI Architecture
Architecture questions raised by AI systems, infrastructure, security, trust, business rules, and enterprise data boundaries.
AI architecture is where questions about systems, data, trust, security, and business process start to overlap.
These posts look past the model layer and ask what has to change around it: infrastructure, APIs, data boundaries, governance, and the practical limits of automation.
AI Is About to Expose Your Security Posture — Not Improve It
Why AI security tools may reveal weak engineering practices faster than they improve them, and what that means for architecture and risk.
AI Didn’t Break Infrastructure — It Exposed Its Edges
How AI exposes the boundaries between infrastructure, data, security, and governance rather than simply creating a new infrastructure problem.
NVIDIA Didn’t Miss the Data Problem — They Drew a Boundary Around It
Why NVIDIA’s AI architecture story is strong on compute and infrastructure, and why the data boundary still needs separate architectural attention.
The Agent at the Gate — Why Agentic Access Breaks the Unwritten Rules of Database Security
The application layer has long been the practical gatekeeper for enterprise data. Agents weaken that assumption and put pressure back on the database boundary.
Who Asked That? Identity, Accountability and the Agentic Query
Agentic access breaks the link between end-user identity, query scope, authorisation and audit. The database needs to know more than a service account.
Data Integrity in an Age of Autonomous Writes
Agents can create plausible, structurally valid, semantically wrong data at scale. Integrity needs controlled write surfaces and database-enforced invariants.
The Return of ACID — Why Agents Make Strong Transactions Non-Negotiable
Sagas and eventual consistency were designed for bounded failure modes. Non-deterministic agents force us to re-examine transaction boundaries.
The Object Storage Default — Why Cloud Architects Get This Wrong
Object storage should not be the default for governed enterprise content. It creates separate security, recovery, audit and consistency boundaries.
The Disconnected Vector Store — Why AI Storage Needs to Move Inside the Boundary
Vector stores and agent memory inherit the governance problem of external storage. AI context should not become an ungoverned parallel truth.