The Scoreboard · Ins7ghts Daily, Friday 20 November 2026

Every Friday this letter grades its own forward calls. The rule since 6 November: the board holds every call whose deadline has passed, from any date, each graded once, and a verdict counts only when a dated event in public reporting meets every qualifier in the call. AI does the reading, the counting and the first pass at the grading. Yves decides what counts and checks what it hands him.

0 right 32 wrong 7 not a real call 0 still open

Three came due since the last board. All three are wrong, and they fail the same way.

13 May, wrong. It needed a named Fortune-500 procurement win where a second-tier cloud and a hyperscaler split an AI workload inside the same tender, with a real cost benchmark cited by the buyer's procurement chief, by 13 November. No such win was reported in the window. Second-tier clouds did keep taking AI work off the hyperscalers, which is the part the call got directionally right, but the evidence it asked for was a buyer publishing its own cost comparison, and buyers do not do that.

14 May, wrong. It needed a named top-five US health insurer, retail bank or wealth manager to ship a branded agent inside the GPT Store with a named compliance disclosure, by 14 November. No such launch exists in the window. Branded assistants did ship in regulated industries, inside the firms' own apps and portals, where the firm controls the disclosure text and can change it on a Tuesday.

The rest of this piece is for subscribers.

Ins7ghts Daily is free. One letter a morning from Yves Mulkers on what moved in data and AI, and why it matters to your side of the table.

Already subscribed? Sign in at the top of the page.