T23

See the scoped evidence_register and qualification_protocol. Structural plausibility, descriptive surveys and laboratory benchmarks do not alone identify a local commercial policy effect.

See the scoped evidence_register and qualification_protocol. Structural plausibility, descriptive surveys and laboratory benchmarks do not alone identify a local commercial policy effect.

Category agent adoption can create congestion, improve relevance, expand demand or change platform power. Net externalities and first-mover decay are empirical questions requiring interference-aware measurement. The model already distinguishes drift, regime change and topology change. Link those distinctions to a transition process influenced by the firm, rivals, platforms and external conditions. A changepoint alarm identifies a statistical change, not its cause. Test whether policy changes, delivery shifts, selection, competitor response or external shocks explain it.

• Filippou, Quach and Jha — Pseudo Incrementality Testing — 16 September 2026. Full preprint. Changepoint detection plus counterfactual forecasting, two GeoLift benchmarks and synthetic stress tests. Only one of three nominal 90% intervals covers truth in the stress test; weak controls yield large bias. Source: https://arxiv.org/html/2609.18257v1
• Martin Bichler — Agentic Markets — 6 June 2026. Full journal article. Conceptual synthesis of agent search, differentiation, market power and platform architecture. Source: https://link.springer.com/article/10.1007/s12525-026-00906-y
• Quanyan Zhu — Agentomics — 9 June 2026. Full preprint. Formal/normative workflow-surplus and cooperative allocation framework. Source: https://arxiv.org/html/2606.14769v1
• Microsoft Research, Bansal, Mirza and colleagues — Red Teaming a Network of Agents — 30 April 2026. Primary report. Internal network of more than 100 agents; malicious-message propagation, borrowed reputation and provenance loss. Source: https://www.microsoft.com/en-us/research/blog/red-teaming-a-network-of-agents-understanding-what-breaks-when-ai-agents-interact-at-scale/
• Yan, Huo, Shang, Peng and Wu — CEO Arena — 28 September 2026. Full preprint. Eight-firm, 500-day stylized competitive environment; matched replacements, 27 main and 26 robustness runs. Private gain can coexist with aggregate market loss. Source: https://arxiv.org/html/2609.34821v1

• No fitted local coefficient or complete-policy deployment evidence is supplied.
• parameter uncertainty not fully propagated, small validation base, no guarantee that pauses are exogenous. Useful refusal conditions rather than a substitute for experimental design. C3/C5/C7.
• no new calibrated market experiment. Lower technical search costs do not guarantee exhaustive search, equal information, higher welfare or a particular equilibrium. C2/C3/C7/C8.
• the security-operations numerical example is hypothetical. Shapley allocation is a proposed normative benchmark, not an observed pricing equilibrium; coalition values themselves require estimation. C5/C7/C8.

Declare eligible population/business kind, target and denominator, pre-action information, policy/comparator/assignment, outcome horizon and coverage, cost/authority/configuration, uncertainty, causal assumptions, falsifier and expiry. Current coefficients/qualification remain unmeasured unless explicitly linked to local evidence.