Fine-Tuning on Enterprise AI Traffic: How to Keep Permission Boundaries Intact
Guides

Fine-Tuning on Enterprise AI Traffic: How to Keep Permission Boundaries Intact

Fine-tuning on enterprise AI logs can leak what some users alone may see. Label, train and serve specialist models within permission boundaries.

Craig Alberino · Nov 20, 2026 · 10 min read
Smartflow for AI Data Centers: More Tenant Work from the Capacity Already Live
Enterprise AI

Smartflow for AI Data Centers: More Tenant Work from the Capacity Already Live

Permits, power and GPUs arrive on their own schedule. Smartflow meters and governs every AI call per tenant, so operators get more work from live capacity.

Craig Alberino · Nov 20, 2026
Whose Balance Sheet Carries AI Capex?
Enterprise AI

Whose Balance Sheet Carries AI Capex?

Demand for compute sets the capex. Rent or own decides whose balance sheet takes it, and how long the demand is committed.

Craig Alberino · Nov 20, 2026
Self-Hosting Is Not Ownership: Governing the Models Inside Your Walls [2026]
Guides

Self-Hosting Is Not Ownership: Governing the Models Inside Your Walls [2026]

Self-hosting decides where a model runs. Owning it takes provenance, limits on what it may touch and a record of what it did. Four cases, four controls.

Craig Alberino · Nov 19, 2026
SR 26-2 Leaves Generative AI Out of Scope: How Banks Govern LLMs and Agents Anyway
Guides

SR 26-2 Leaves Generative AI Out of Scope: How Banks Govern LLMs and Agents Anyway

SR 26-2 replaced SR 11-7 and left generative and agentic AI out of scope. How banks govern LLMs and agents anyway, practice by practice, and what to keep.

Craig Alberino · Nov 17, 2026
HTTP 402 for AI Agents: Turning Agent Traffic into Paying Customers
Guides

HTTP 402 for AI Agents: Turning Agent Traffic into Paying Customers

HTTP 402 lets a publisher answer an AI agent with an offer instead of a block. What RFC 9110 says, how x402 and pay per crawl use it, and what to require.

Craig Alberino · Nov 13, 2026
MCP Security: How to Govern the Tool Path for AI Agents [2026]
Guides

MCP Security: How to Govern the Tool Path for AI Agents [2026]

MCP security for enterprises: tool poisoning, rug pulls and token passthrough, what the MCP spec requires, and eight controls for your agents' tool path.

Craig Alberino · Nov 12, 2026
Building Expert Systems with Celonis: Context, Authority and the Record
Guides

Building Expert Systems with Celonis: Context, Authority and the Record

Celonis supplies the knowledge base. An expert system that acts also needs a model chosen per call, a named person's authority and a record kept outside the agent. Seven steps.

Craig Alberino · Nov 11, 2026
Rent or Own: How to Decide Where Each AI Workload Runs [2026]
Guides

Rent or Own: How to Decide Where Each AI Workload Runs [2026]

Self-hosted LLM or API? Decide workload by workload: the data it touches, the capability it needs, cost at volume, stability. Govern both under one policy.

Craig Alberino · Nov 10, 2026
AI Agents in Vendor Security Reviews: Why the Transcript Cannot Be the Record
Guides

AI Agents in Vendor Security Reviews: Why the Transcript Cannot Be the Record

When two companies' AI agents run a vendor security review, the transcript is no record. What a refereed, signed record of checked answers requires.

Craig Alberino · Nov 6, 2026
Step-Up Approval for AI Agents: A Named Human on Every High-Risk Action
Guides

Step-Up Approval for AI Agents: A Named Human on Every High-Risk Action

Step-up approval pauses an AI agent's high-risk action until the person whose authority it uses approves. How it works and which actions need it.

Craig Alberino · Nov 5, 2026
How to Qualify a New AI Model Before It Takes Production Traffic: Shadow, Compare, Canary, Promote
Guides

How to Qualify a New AI Model Before It Takes Production Traffic: Shadow, Compare, Canary, Promote

How to qualify a new AI model before it takes production traffic: eligibility, shadow testing, held-out comparison, canary, promotion and rollback.

Craig Alberino · Nov 3, 2026