Let's be clear: agentic AI is not a magic intern quietly cleaning up your codebase; it is more like a new hire with root access and no employee handbook. If your IT strategy does not start with trust, guardrails, and clear accountability, agentic tools will amplify risk faster than they amplify productivity.
Recently, I was on the #AITalk Podcast with Business Reporter, discussing this very topic with host Kevin Craine, Bryan Finster, Principal Engineer, and Derek Ashmore, AI Enablement Principal at Asperitas Consulting.
The interesting shift is that developers move from typing code to defining outcomes and specifications, but that only works if your governance and delivery pipelines are equally modern. Quality has to move left into the model and workflow design, with automated tests, static and dynamic analysis catching what human reviewers and traditional QA would miss. The smartest early use cases are low-risk and high-leverage, such as agents that keep documentation up to date while critical changes in production still go through human review. And instead of waiting for regulators, pay attention to defect patterns, customer security questionnaires, and real-world incident news to decide where your own red lines should be.
TL;DR
🧠 Start with trust, guardrails, and accountability
⚡ Keep humans in the loop for critical changes and production code
🎓 Shift quality left with tests, plus static and dynamic analysis
🔍 Use low-risk agents first, like documentation and onboarding, and build your own controls before regulation arrives
#AgenticAI #SoftwareEngineering #DevOps #ITStrategy #security #privacy #cloud #infosec #cybersecurity
https://www.business-reporter.co.uk/ai--automation/ai-talk-updating-your-it-strategy-for-agentic-ai
Lawondyss
Brian Greenberg 
The Internet is Crack
Andrew McGivern
Joaquim Homrighausen
4sphere