Curious where will AI slide generation stop providing value?

We've spent the past year building Verso and watching how strategy teams actually adopt AI in their slide workflows. Here's our honest forecast.

~70% capability (now): Mechanical production automates. Formatting, layout consistency, bulk changes. The analyst skill premium on "fast in PowerPoint" drops; the premium on knowing what a slide should communicate rises.

~85% (18-24 months): Complex elements - charts, tables, multi-column layouts - become reliably editable. The feedback loop between senior judgment and slide output collapses from hours to minutes. More iterations, higher quality ceiling.

~95%+ (speculative): The static file format itself may evolve - decks that regenerate from structured inputs rather than being manually assembled. We're genuinely uncertain whether this is better. The static deck has value as a shareable artifact, a specific argument at a specific moment.

What doesn't change: the analytical work, the judgment, the client relationship.

Full piece: get-verso.ai/blog/future-of-slides-ai

Curious what others building in this space think - especially on the 95% scenario. Worth preserving the static format, or does it give way to something dynamic?