Kesem

Kesem is an AI presentation tool that turns a topic prompt into a structured storyline, then builds slides with generated charts, file-based analysis and Unsplash visuals. Decks export to PDF or PPTX on paid plans.

Kesem is a web-based presentation builder that starts from a written topic rather than a blank canvas. You type the subject you want to present, and the tool returns a set of candidate ideas or subtopics. Each idea you keep becomes roughly one slide, so choosing eight ideas yields a deck of at least eight slides. From there Kesem assembles the narrative itself, then lets you adjust the wording and styling, swap imagery, and export. The feature set leans toward data-driven decks. Beyond text slides, Kesem generates charts from instructions and data you supply, and it can ingest uploaded files to produce a fuller analysis: questions, charts, insights and action items drawn from your material. That combination makes it most relevant to analysts, consultants, client-facing professionals and anyone who has to turn a spreadsheet or report into something a room full of stakeholders can follow. Multi-language output supports decks aimed at non-English audiences, and the export step covers PDF and PPTX so finished work can move into standard presentation software. In practice the workflow is three stages: define the core message with a prompt, select the subtopics that shape the storyline, then customize text and visuals. Uploaded files and prompts feed the analysis features, and downloaded data is available on the entry plan. Image sourcing pulls from Unsplash, with attribution links attached to each picture, and images can be replaced with your own. Kesem sits among AI deck generators, but positions itself around narrative structure and file-based analysis rather than template variety. The catalogue of designs is limited (the site points to layouts known from Canva) and custom company templates are described as a future goal rather than a shipped feature. The FAQ states the product runs on GPT 3.5 with knowledge only to September 2021, so current facts and recent data need to come from your own uploads. Downloading is gated behind a paid membership.