Notifications
Clear all
Topic starter
19/08/2026 4:03 pm
TL;DR: TrueFoundry’s analysis argues that an agent economy is emerging around software agents that hold budgets, discover counterparties, and transact at machine speed, but the real constraint is identity, trust, and payment infrastructure, not model quality, according to TruFoundry. The post’s core implication is that governance assumptions built for human-paced access and static principals break once agents become economic actors.
NHIMG editorial — based on content published by TruFoundry: Agent Economics, No. 1: What Is the Agent Economy and Who Gets to Design It?
By the numbers:
- 33% of organisations report their AI agents have accessed inappropriate or sensitive data beyond their intended scope.
Questions worth separating out
Q: How should security teams govern agent budgets in an AI economy?
A: Security teams should treat budget as a privileged entitlement with explicit ownership, thresholds, and revocation rules.
Q: Why do agent economies complicate IAM and PAM models?
A: They complicate both because identity is no longer only about access to systems.
Q: What breaks when agent transactions outrun human review cycles?
A: Periodic review breaks because the relevant authority may exist only briefly and may be exercised many times before a human can inspect it.
Practitioner guidance
- Define agent budget as a governed identity attribute Treat spend limits as part of the entitlement model, with explicit owners, review cadence, and revocation triggers for every agent principal.
- Separate discovery from transacting authority Do not allow every agent that can find counterparties to also execute purchases, tool calls, or settlement actions without a second control boundary.
- Instrument machine-speed transaction logs Ensure logs preserve principal, budget, counterpart, and policy state so investigations can reconstruct what the agent did and why it was allowed.
What's in the full article
TruFoundry's full blog post covers the analytical detail this post intentionally leaves for the source:
- The step-by-step derivation of the agent-economy definition and the toy permeability metric used in the article.
- The comparison between intentional and emergent market design, including how the article frames sandbox economics.
- The article's discussion of agent identity, trust services, and payment rails as the missing infrastructure layers.
- The editorial ledger of open questions the series plans to revisit as evidence accumulates.
👉 Read TruFoundry's analysis of the agent economy and identity rails →
Agent economies: what identity teams need to design for now?
Explore further
19/08/2026 4:38 pm
The identity layer becomes market infrastructure once agents can transact. Traditional IAM treats identity as a gate to applications and data. In an agent economy, identity also determines whether an agent can discover counterparties, establish trust, and carry delegated budget. That shift makes identity governance part of market design, not just access administration. Practitioners should read this as a structural change in what identity is for.
A few things that frame the scale:
- 92% agree governing AI agents is critical to enterprise security, yet only 44% have implemented any policies to do so, according to AI Agents: The New Attack Surface report.
- Only 52% of companies can track and audit the data their AI agents access, leaving 48% with a complete blind spot for compliance and breach investigation.
A question worth separating out:
Q: How do organisations decide whether to trust agent-to-agent exchange?
A: They should base trust on verifiable identity, policy-bound budgets, and transaction logging that makes the counterparty relationship auditable. Trust cannot be inferred from model sophistication or gateway placement. It has to be anchored in governance that explains who the agent represents and what it may do.
👉 Read our full editorial: Agent economies expose the identity layer as market infrastructure
