Side Stage lead line. "Everyone has the same AI. So we made one that lives in your building." Recommended: yes. It's the one statistic as a sentence, and it's how you say it.
Infosys lead line. "Selma, working inside the building. Three organisations live, forty asking." Recommended: change "Three organisations live" to "Two live, one signing" unless RCSI is installed (see below). Otherwise keep.
The closing ask. Project rule 4 says no closing ask slide; the Side Stage deck ends on the ask. Recommended: no closing ask. End on the tagline, with the ask as a beat in the middle, the way your own decks run.
Personal-Selma lines in three deep dives. The investor rules say leave the home long game out. Recommended: cut them from generated decks (the home deep dive, the donated unit line, "then homes", "a personal Selma for homes").
"Three organisations. Live now." RCSI's charter is ready to sign, not installed. Recommended: "Two live, one signing." True today and still a good line.
Lead or co-invest for Side Stage? No recommendation; your call and the deck changes with it.
Elli's three 2025 pieces. Medium and the AFR blocked the window. Recommended: you read them before the meeting, or paste the text into the Investadeck folder as inputs/elli-2025.md and the strategist folds them in.
INV-007 landing page, done looks like and check. Proposed: one page at investadeck.brock.cx that tells a founder what Investadeck does, shows the Side Stage deck as the example, with the price line as a placeholder. Check: gate, one inspection at 1440 and 390, the reviewer. Recommended: yes.
Synthetic imagery. Ten scene images and both videos are renders, unlabelled; two show a New York skyline. Recommended: label them synthetic in the master now, and replace the two skylines with a real photo of a pilot unit when one exists.
Launch wording. The page says "paid launch"; the project trap says investor material says "commercial launch". Recommended: commercial launch.
Accuracy line. The accuracy deep dive (7 Oct: 7,754 questions, 43 of 49, 360 refusals) retires "ten thousand questions, one invented figure". Recommended: retire it everywhere, update the Selma skill, and keep accuracy out of the lead as you've said.
RCSI by name. Not on the 21 Jul nameable list. Recommended: don't name RCSI in decks until you confirm.
Pricing. The page says $150K over 3 years ($100K not-for-profit); the Selma skill says setup plus per seat, $30K to $150K a year. Which is current?
Sectors. "14 sectors asking" from the page's own list; the Selma skill says 18. Which count, and from where?
The 57 percent. "57% say so" in the problem beat's presenter note has no source. Recommended: drop it until sourced.
25 deals. The ask commits to 25 deals in the first half of 2027; deck-facts ruled out client-count commitments. Recommended: drop the number, keep the shape ("the first wave of installs").
Disk on the Mac. Done: Simon freed space by hand, 1 Nov. Nothing to decide.
Yes, no or maybe on each part. Comment where it matters. The studio pulls these before it builds.