• By Nick Patel
  • 14 Nov, 2026
  • Industry

Can AI Write an 80-Episode Micro-Drama? Screenburn’s Writers Room vs the Vertical 52

By Nirav Patel, Founder, Vertical Charts

An 80-episode micro-drama season used to take a writers’ room weeks. Screenburn Group says its new software can do it in a few hours. The UK scripted-entertainment company launched The Writers Room this week, a generative AI tool that takes a writer’s outline and returns a full 80-episode vertical season, roughly 29,000 words of screenplay, according to Broadcast’s November 12 report.

We cover the shows people actually watch, so we wanted to check the pitch against our own data. How does an 80-episode template line up with the hits on this week’s Vertical 52? And can a script score predict what an audience does after release? Here’s what we found.

What Screenburn is selling

According to Broadcast, a writer plans a story on Screenburn’s screenwriting platform: a title, a one-line summary, the characters and answers to a set of questions about them. Then the Writers Room writes all 80 episodes. The company says it writes to rules for how a season opens, turns and ends, drawn from data on more than 100,000 vertical series.

A second tool, the Retention Engine, then reads every episode and scores it out of 100 on four things: how episode one opens, how each episode ends, how often each episode tells the viewer something new, and how tension builds across the season. A score of 80 or higher earns a “Screenburn Certified” mark. Screenburn says it filed a US patent application for the engine in August, has more than 20 scored seasons on sale (the highest scored 97), and charges US$1,350 a month on a 12-month contract for a writer plus up to five colleagues, including one season a month.

CEO Dr Charles Kriel’s argument is about cost. Anyone has been able to make AI pictures for a year, he told Broadcast, but nobody could write the show in a couple of hours. Now a writer can have a full, scored season “before anyone spends a dollar making it.”

Our chart’s hits are shorter than 80 episodes

The first thing that jumps out is the number 80. It’s a reasonable industry default, and Broadcast describes micro-drama seasons as running 60 to 100 episodes. But the shows leading our chart right now mostly run shorter.

We have verified episode counts, taken from each show’s official app page, for 24 of the 52 titles on the November 12 Vertical 52. Their median is 56 episodes. Only three run 75 or longer, and just one passes 80: DramaBox’s long-running Fortunes Unveiled: My Husband Is a Big Shot (89 episodes, #24). Fourteen of the 24 have 60 episodes or fewer.

The top of the chart is shorter still. For the seven top-10 shows with a count, the median is 52 episodes:

This matches what we found in our ReelShort vs DramaBox episode-count analysis in October: DramaBox’s charting shows had a median of 52 episodes against ReelShort’s 63, and DramaBox’s newest releases were getting even shorter. Since DramaBox holds nine of this week’s top 10 seats, an 80-episode season isn’t the shape of the current winners.

That matters for money as well as storytelling. As we showed in our look at what it costs to finish a micro-drama, every extra paid episode is another unlock a viewer has to accept. Longer isn’t automatically better for the viewer or the app. A tool built around 80 episodes is only useful if writers can easily dial it down to 45 or 55.

A script score isn’t an audience score

The Retention Engine is the most interesting part of the launch, and also the easiest to over-read. It scores a script before anything is shot. Our VertiScore works from the other end: it measures what real audiences and the apps do after release, using velocity (how fast a show is adding viewers), reach, placement (how prominently the app features it) and save rate.

Two things on this week’s chart show why a script score can only tell part of the story.

Length doesn’t predict saves. The highest save-rate score in our top 10 belongs to the longest show there, I Wish It Were You (78 episodes, save-rate component 96.9). But the #1 show, at 43 episodes, scores 88.4 on the same measure. A good hook-and-cliffhanger structure clearly works at both lengths, which is good news for Screenburn’s rule-based approach, as long as length is flexible.

Placement can outweigh the writing. In our November 12 recap, My Sister Is the Warlord Queen fell 18 places in one week while still having the most views on the chart. Its script didn’t change. ReelShort just stopped featuring it. Meanwhile The Lost Quarterback Returns jumped 31 places when it returned to the featured rails. No script score can see an app’s merchandising calendar.

AI writing vs AI pictures

Screenburn is also making a narrower bet than most AI micro-drama news. The AI wave so far has mostly been about images. Beijing Review reported on November 10, citing the China Netcasting Services Association, that about 122,000 of the roughly 128,000 micro-dramas released in China in Q1 2026 were AI-generated. The complaints it lists are mostly visual and logical: faces changing between shots, props switching mid-scene, characters showering fully clothed.

On our chart, AI-labeled shows are still a minority. Four of ReelShort’s 26 titles carry ReelShort’s AI label this week: Rejecting My Five Female Mates (#18), The Great and Powerful Genie (#19), I Accidentally Seduced the Siren King (#45) and My Cursed Merman Husband Rules the Deep (#49). That’s down from five a few weeks ago in our AI vs live-action analysis. A writing tool sidesteps the visual-continuity problem entirely, since the output can be shot with real actors. That could make it more useful to the live-action studios that still hold most of the chart.

Screenburn isn’t alone in turning production tooling into a product. The Economic Times reported this week that Peak XV-backed Dashverse shut its consumer app DashReels and is raising money mainly for Frameo AI, its business video software. EloElo, which just raised $80 million in a round led by Hornbill Capital and NBCUniversal, is also investing in its own AI creative suite, Dolphin AI.

Our take

Faster development is real value. A producer who can read a full season, scored, in an afternoon can kill weak ideas cheaply. But the chart suggests three caveats. The winning seasons right now are closer to 50 episodes than 80. Retention on a page isn’t retention on a phone. And a show’s fate often depends as much on whether the app features it as on what’s in the script. If Screenburn-written shows start reaching the apps, we’ll track them on the Vertical 52 and see whether a 90-plus script score turns into a high VertiScore. That will be the real test.

Data note: episode counts come from official ReelShort and DramaBox title pages for 24 of the 52 titles on the November 12, 2026 Vertical 52. Screenburn product details are as reported by Broadcast; we have not tested the software.

Tags: