News

IRONSTONE WEEKLY

One signal. Why it matters. What to watch next.

LAST WEEK: OCTOBER 5–11, 2026

Three signals from AI, enterprise software and insurance.

THE UNCERTAIN — AI RISK IS BECOMING A COVERAGE ARGUMENT

News: Marsh warned the insurance market against broad or absolute AI and generative-AI exclusions.

Why it is interesting: If AI is excluded as a category, ordinary digital operations can become uninsurable. The better question is not simply whether AI was involved. It is what the system did, under whose authority, and what evidence exists.

Read: Marsh warns market against AI exclusions 

THE GOOD — THE SEMANTIC LAYER IS BECOMING THE PRODUCT

News: Infor introduced an Industry AI architecture in which agents coordinate through a shared semantic layer and a catalogue of more than 350 use cases.

Why it is interesting: Agents without shared meaning become disconnected assistants. A semantic layer gives them a common model of the business—the difference between a collection of tools and a working system.

Read: Infor unveils its expanded Industry AI architecture 

THE HARD PROBLEM — AN AGENT NEEDS AN IDENTITY

News: RSA announced Agent ID for regulated industries, focused on registering agents and proving the authority behind consequential actions.

Why it is interesting: Identity is becoming the control plane for agentic systems. If an organization cannot name an agent, constrain its authority and reconstruct what it did, governance is only a policy document.

Read: RSA Agent ID closes the agentic identity gap 

THE SIGNAL

The pattern is not simply “more AI.” It is a new control plane: shared meaning, named authority and insurable boundaries. That is the thread I will keep watching.

The public preview stops here. Paid subscribers get the complete weekly issue and working notes.

Paid edition: Read and subscribe on Substack 

Prefer a conversation? Email rob@ironstoneadvisory.com or book a 30-minute meeting.