AI mediation of economic decisions: eight registered forecasts to 2031

EA-MEDIATED-SHARE-01 registers eight forecasts on AI-mediated commerce, each with a check date. They cover AI-referred dollar share at the end of 2026, agent-executed share in 2030, the scope and revision of the McKinsey and Bain headline figures, professional-services citation rate, machine-readability spread, accuracy reporting and the comparability of platform denominators. One unbuilt test designs a randomised inscription trial scored for accuracy rather than presence.

8
forecasts
1
unbuilt test
1
deposits
2026-09
earliest

Each claim below is registered at the earliest deposit that states it, and quoted verbatim from the deposit text, with every later statement, its dated revisions, and the later records that claim it came true or failed. The claims are the deposits' own; this page records them and does not judge them.

PRED-1632.A1forecastregistered 2026-09-20 in #1632 · AI Mediation of Economic Decisions: A Registered Forecast, a Conditional Valuation, and a Statement of Burden
1. AI-referred dollar share, end of 2026. Requires AI-referred session share and relative average order value, or directly reported revenue share, for a period ending Q4 2026. Conversion differential alone will not settle it. Check 2027-03-31. Claim: 3 to 8 percent at v = 1, restated if v is published.
falsifier
If AI-referred dollar share at end of 2026 is below 1 percent, the retail channel is further out than this assumes. It would not establish a ten-year delay, and it would say nothing about mediation in fields these instruments do not observe.
window Check 2027-03-31
series: EA-MEDIATED-SHARE-01 registered forecasts
PRED-1632.A2forecastregistered 2026-09-20 in #1632 · AI Mediation of Economic Decisions: A Registered Forecast, a Conditional Valuation, and a Statement of Burden
2. Agent-executed share, 2030. Census e-commerce totals for 2030 against any published measurement of agent-executed transaction value with a stated definition. Check 2031-06-30. Claim: 5 to 15 percent. If no such figure with a stated definition exists, the item resolves UNRESOLVED — the outcome this document most expects.
falsifier
If McKinsey's trajectory proves closer than Bain's, the deployment-constraint reading is wrong and mediation arrives faster than section 4 allows.
window Check 2031-06-30
PRED-1632.A3forecastregistered 2026-09-20 in #1632 · AI Mediation of Economic Decisions: A Registered Forecast, a Conditional Valuation, and a Statement of Burden
3. Forecast scope. Claim: McKinsey does not publish a scope definition for its orchestrated-revenue figure permitting comparison with Bain's, before 2028-12-31.
window before 2028-12-31
series: EA-MEDIATED-SHARE-01 registered forecasts
PRED-1632.A4forecastregistered 2026-09-20 in #1632 · AI Mediation of Economic Decisions: A Registered Forecast, a Conditional Valuation, and a Statement of Burden
4. Forecast revision. Claim: at least one of the two headline figures is revised by more than 40 percent or withdrawn before 2029-12-31.
window before 2029-12-31
series: EA-MEDIATED-SHARE-01 registered forecasts
PRED-1632.A5forecastregistered 2026-09-20 in #1632 · AI Mediation of Economic Decisions: A Registered Forecast, a Conditional Valuation, and a Statement of Burden
5. Professional-services citation rate. Claim: rises above 10 percent by 2027-12-31 (Similarweb category series). This resolves a citation-rate prediction and nothing else. It does not establish presence or absence of a market: citation is not the whole channel, and a small number of consequential encounters can carry value.
window by 2027-12-31
series: EA-MEDIATED-SHARE-01 registered forecasts
PRED-1632.A6forecastregistered 2026-09-20 in #1632 · AI Mediation of Economic Decisions: A Registered Forecast, a Conditional Valuation, and a Statement of Burden
6. Machine-readability spread. Claim: Adobe's best-to-worst spread stays above 14 points through 2027-12-31.
window through 2027-12-31
series: EA-MEDIATED-SHARE-01 registered forecasts
PRED-1632.A7forecastregistered 2026-09-20 in #1632 · AI Mediation of Economic Decisions: A Registered Forecast, a Conditional Valuation, and a Statement of Burden
7. Accuracy reporting. Claim: no vendor or academic benchmark publishes entity-resolution accuracy by field, as distinct from presence, mention or citation rate, before 2027-12-31.
falsifier
If a vendor publishes accuracy-by-field with a stated rubric and an open sample, section 8's burden claim weakens immediately and should be withdrawn to the extent that publication covers.
window before 2027-12-31
series: EA-MEDIATED-SHARE-01 registered forecasts
PRED-1632.A8forecastregistered 2026-09-20 in #1632 · AI Mediation of Economic Decisions: A Registered Forecast, a Conditional Valuation, and a Statement of Burden
8. Comparability. Claim: at a re-audit on 2027-09-10, no three of the nine platforms publish mutually reconcilable denominators.
window at a re-audit on 2027-09-10
PRED-1632.F1unbuilt testregistered 2026-09-20 in #1632 · AI Mediation of Economic Decisions: A Registered Forecast, a Conditional Valuation, and a Statement of Burden
Design, stated so it can be criticised before it runs: entities matched on field and size; randomised to inscription against conventional structured data, not against nothing, because a buyer's real choice is against the alternative; fixed query sets issued before treatment across at least three surfaces, since only 11 percent of domains are cited on two; outcomes scored for accuracy on the declared rubric rather than presence; commercial events tracked to something the firm can see; failures recorded and published with successes. That design is also what would put a number on p, which is the term nobody has.