You gotta put more work into the organization on the screen for the reader, not just text bubbles.
Look at comics and manga, pages are split up into panels and they are organized with a directional flow for the reader's eye. This helps set up close up shots, action shots, sets up locations and mood. Doing all this with generative AI is possible - I've seen it done, and I've seen it posted here, but it is going to take a lot more work from YOU to not only get the AI to spit out what you need, but YOU'RE going to be the one to organize the panels on each page, not the AI.