Explore the publication
Author: Enterprise Tech Logic
Explore this topic
Data & AI · Practical guide
Human approval for AI must control the action that actually executes
Place approval before a consequential AI action, bind it to the exact proposal and test changed inputs, expired consent, retries and emergency stops.
IT Operations · Practical guide
Incident review actions should change the next failure
Build a reliability backlog from incident findings, with actions that have an owner, a reason to take priority and a test of the intended improvement.
Data & AI · Practical guide
Data deletion is a workflow across copies, holds and restores
Turn an approved retention policy into a deletion workflow that covers exports, legal holds, backups and restored data, with clear evidence of what remains.
Cloud & Infrastructure · Practical guide
A cloud exit plan should prove the service works somewhere else
Test exports, replacement workflows, operating responsibilities and cutover before renewal. Free data transfer does not make a cloud service portable.
Data & AI · Decision guide
Enterprise AI data controls: approve the feature, not just the provider
Training restrictions, retention and customer logging answer different questions. Check the actual AI service, enabled features and configuration before approval.
IT Operations · Analysis
Platform engineering metrics should follow the developer’s whole task
Faster service creation can hide support queues and rework. Measure developer independence, usable outcomes and reliability across the whole journey.
Data & AI · Decision guide
AI data readiness includes the document you have just revoked
Prepare enterprise retrieval around permission changes, freshness and deletion. Test every derived copy, not just the quality of the source documents.
Cloud & Infrastructure · Decision guide
Cloud repatriation: price the operating model you will inherit
A repatriation case needs more than cheaper servers. Compare transition costs, support, spare capacity and service obligations over the same horizon.
Data & AI · Explainer
RAG or fine-tuning? Start with what the AI gets wrong
RAG supplies information at response time; fine-tuning adapts model behaviour through training. The right choice starts with the failure you need to fix.
IT Leadership · Analysis
When a technology pilot has earned another phase
Decide whether to stop, change or extend a technology pilot using the evidence it can still produce, the incremental cost and a clear decision threshold.
Connected topics
