How I Use Automatic Video Editing AI to Cut Editing Time in Half
How I Use Automatic Video Editing AI to Cut Editing Time in Half
Last updated: July 2026
Author: Greg Preece — I test AI video tools hands-on to help creators and marketers streamline their post-production workflows and edit faster.
If you are still scrubbing through timelines to manually cut out silences, mistakes, and filler words, you are losing hours of your life to tedious tasks. In this guide, I will show you how to use automatic video editing AI to turn a raw, unedited talking-head video into a polished, publish-ready cut in seconds. Here is the exact workflow I use to automate my post-production using Descript's AI co-editor, Underlord.
TL;DR: The Automatic Video Editing Workflow
- The Tool: [Descript] (using the built-in Underlord AI assistant).
- The Workflow: Import your footage, generate a transcription, and let the AI automatically correct your gaze, remove pauses, erase filler words, delete repeated takes, and apply subtitles.
- The Big Time-Saver: The Edit for Clarity feature combines multiple cleanup steps into a single click, instantly cutting a raw 10-minute video down to a tight, 4-minute edit.
- The Benefit: No timeline-scrubbing required. If you can edit a Word document, you can edit your video.
Prefer to watch? Here's the video. Prefer to skim? The full breakdown is below.
Quick Link
Try it here: Try Descript →
Table of Contents
- Why Traditional Video Editing Is Broken
- The Core Solution: Descript and Underlord AI
- Step-by-Step: The 6 AI Features That Automate Your Cuts
- How Much Time Does AI Actually Save? My Testing Results
- Frequently Asked Questions
Why Traditional Video Editing Is Broken
Traditional editing in software like Premiere Pro or DaVinci Resolve requires you to look at audio waveforms, manually slice out dead space, find every "um" and "uh," and delete mistake takes one by one. For a 30-minute recording, this basic cleanup process can easily eat up an hour or more of your time before you even start adding B-roll, music, or graphics.
By utilizing an ai automatic video editor, you can offload these repetitive, manual tasks entirely. Instead of focusing on technical cuts, you can direct your energy into scripting and recording your content.
The Core Solution: Descript and Underlord AI
To automate this workflow, I use Descript. Descript works by transcribing your video file instantly upon upload, allowing you to perform text-based video editing.
Most of its automatic editing power is housed within Underlord—Descript's AI co-editor. Underlord is a suite of AI-driven tools that can read your transcript, analyze your audio waveform, and make edit decisions for you.
Below, I’ll walk you through the six core automatic features I use, showing you exactly how they perform in a real-world edit.
Step-by-Step: The 6 AI Features That Automate Your Cuts
1. Eye Contact Correction
If you use a teleprompter or read your script off a screen to the side of your lens, your eyes will naturally drift away from the camera. This makes you look less engaged with your audience.
In the past, fixing this meant either re-recording the segment or covering up your gaze shifts with B-roll. Descript's Eye Contact AI fixes this in a single click.
Caption: Underlord's Eye Contact tool corrects gaze direction instantly, making it look as though you are looking directly at the camera.
How to use it:
- Drag and drop your video file into Descript.
- Click on the Underlord (sparkle) icon in the sidebar.
- Select Eye Contact.
- The AI will analyze the video and realign your pupils so you are looking directly at the lens.
In my test, I was reading a script off-camera for the entire recording. After running the tool, the AI successfully fixed my eye contact for the entire duration of the video in just a few seconds.
2. Shorten Word Gaps
Every talking-head video has pauses where the speaker is thinking, reading their notes, or resetting for the next sentence. Finding and cutting these silences manually takes a massive amount of time.
Using Descript’s automatic silence remover cuts these out globally.
Caption: The Shorten Word Gaps menu showing 55 silences detected in the raw footage, ready to be collapsed in one click.
How to use it:
- Open the Underlord panel.
- Select Shorten Word Gaps.
- Set your parameters. For my edit, I set it to find any silence longer than 1 second and shorten it to 0 seconds (completely removing the pause).
- The AI found 55 individual silences in my 10-minute raw footage.
- Click Shorten All.
This action instantly trimmed my 10-minute video down to 5 minutes and 50 seconds, cutting out over 4 minutes of dead air in about 20 seconds.
3. Text-Based Editing
While not strictly an "automatic" one-click action, Descript's fundamental design relies on text-based editing. When you import a video, Descript transcribes your audio word-for-word on the left side of your screen.
Caption: Deleting words in the text transcript instantly deletes the corresponding video and audio frames on the timeline.
If you want to delete a sentence, a phrase, or even a single word, you don't need to cut the waveform. You simply highlight the text on your screen and press Backspace. The video timeline updates automatically, creating a clean cut. It makes video editing as simple as editing a Google Doc.
4. Remove Filler Words
Filler words like "um," "er," "uh," and "hmm" make speakers sound less confident and slow down the pacing of a video.
Manually finding these short sounds on a timeline is incredibly tedious. Descript allows you to remove filler words automatically.
Caption: The filler words removal interface showing 12 flagged verbal fillers across the timeline.
How to use it:
- Open Underlord and click Remove Filler Words.
- The AI scans your entire transcript and flags verbal filler words.
- In my test, it flagged 12 filler words.
- Click Remove All.
This single click cleaned up my delivery and brought the video length down from 5 minutes and 47 seconds to 5 minutes and 33 seconds.
5. Remove Retakes
When recording unscripted or scripted videos, you will occasionally make mistakes, stumble over words, and repeat sentences until you deliver them correctly.
Typically, you have to find each failed attempt, cut it out, and align the good take. The Remove Retakes tool uses AI to recognize when you've repeated yourself, automatically keeping only your final, best delivery.
Caption: Underlord identifying and striking out redundant takes, leaving only the clean, final delivery intact.
How to use it:
- Click Underlord and select Remove Retakes.
- Confirm the selection.
- The AI reviews the patterns in your speech (for instance, if you repeat "sound enhancement" three times before completing the sentence).
- It strikes through the incorrect drafts and keeps the correct version.
In my session, this dropped the timeline further from 5 minutes and 33 seconds down to 5 minutes and 5 seconds, completely eliminating false starts.
6. Edit for Clarity
If you want to skip doing all the steps above individually, you can use Descript's most powerful combined feature: Edit for Clarity.
This feature acts as a comprehensive, automatic editor. It removes filler words, cuts word gaps, removes bad retakes, and identifies unnecessary or off-topic tangents that ruin the flow of your content.
Caption: Setting the Edit for Clarity slider to "Heavily Intense" to execute a tight, fast-paced cut.
To test this, I undid all prior changes to return to my raw, 10-minute video.
How to use it:
- Click Underlord and choose Edit for Clarity.
- Adjust the slider to set the intensity of the cuts. I set mine to Heavily Intense to maximize the edits.
- Click Submit.
Rather than leaving me with a 5-minute video (which is what I got when running the individual tools), Edit for Clarity delivered a super tight 4-minute and 9-second cut. This is because the AI went beyond basic silences and fillers—it actually identified fluff and off-topic sentences in the script and trimmed them down for a punchier result.
7. Automatic Captions
Adding dynamic captions can increase your video's viewer engagement significantly. However, transcribing and syncing text word-for-word takes hours if done manually.
Because Descript has already built a highly accurate text transcript of your video, turning it into on-screen subtitles takes zero extra manual effort.
Caption: Applying dynamic, word-level highlighted captions from Descript's style templates.
How to use it:
- Click the Captions icon in the top menu bar.
- Select your preferred subtitle template style from the list.
- The AI instantly generates fully synchronized captions across your entire timeline.
How Much Time Does AI Actually Save? My Testing Results
To show you exactly how much fat can be trimmed using an ai video editor, here are the timeline results from my raw 10-minute clip:
| Editing Step | Video Duration | What was Trimmed |
|---|---|---|
| Raw Footage | 10 mins 7 secs | Baseline footage with pauses, stumbles, and script reading |
| Shorten Word Gaps | 5 mins 50 secs | 55 silences (> 1.0 second) |
| Remove Filler Words | 5 mins 33 secs | 12 instances of "um," "er," and "hmm" |
| Remove Retakes | 5 mins 5 secs | Mistakes, stumbles, and false starts |
| Edit for Clarity (Single Pass) | 4 mins 9 secs | Combined gaps, filler words, retakes, and redundant script fluff |
Using these AI features didn't just shorten the video; it transformed a slow, choppy raw recording into a fast-paced, highly engaging edit. Best of all, instead of spending 30 to 60 minutes slicing waveforms, the entire automated cleanup process took under a minute.
If you want to speed up your editing workflow, remove mistakes automatically, and publish polished content in a fraction of the time, try utilizing an automated AI workflow yourself.
Try it here: Try Descript →
Frequently Asked Questions
Can I manually adjust the AI's automatic cuts?
Yes. Descript’s text-based editor operates non-destructively. Any cut made by Underlord (such as a word gap, filler word, or retake) is simply struck through in your transcript sidebar. If the AI accidentally cuts something you want to keep, you can highlight the struck-through text and restore it instantly.
Does the Eye Contact AI look realistic?
Yes. Descript's Eye Contact feature uses generative AI to shift your pupils so they look directly at the lens. It does this while preserving your natural facial movements and blinking patterns, making it look seamless, even if you were reading a script off-camera.
When should I use "Edit for Clarity" instead of individual tools?
Use Edit for Clarity when you want a rapid, tight rough cut of raw talking-head footage. If you prefer a more conservative edit where you only remove dead pauses and specific filler words while keeping every word of your script, it is better to use Shorten Word Gaps and Remove Filler Words individually.