A Week in the Life: Building a Multi-Platform Video Workflow with AI Tools

# A Week in the Life: Building a Multi-Platform Video Workflow with AI Tools

Date: 2026-07-07

Author: Alex Rivera

Author Role: Video Workflow Engineer & Content Creator

Category: "Video Workflow"

Read time: 8 minutes

Tags: ["video workflow", "AI video editing", "multi-platform publishing", "content creation diary", "video tools"]

Day 1: Mon, July 1 - Planning the Workflow: Tool Overload and Decision Fatigue

Let's start at the beginning: I needed to build a workflow that takes raw footage and spits out optimized videos for YouTube (long-form), TikTok/Reels (short-form), and LinkedIn (professional snippets). No pressure, right?

First step: tool selection. I spent 3 hours scrolling through Reddit threads, watching "best video tools 2026" videos, and testing free trials. The goal? Balance power, cost, and ease of use. Here's how I landed:

- Recording: OBS Studio 30.0.1. It's free, open-source, and I've used it for years--but I considered Streamlabs (too bloated) and Ecamm Live (Mac-only, and I'm on Windows). OBS won for flexibility.

- Main Editing: DaVinci Resolve 19.1. I debated Adobe Premiere (subscription fatigue) and Final Cut Pro (again, Mac). Resolve's free version has 90% of the features I need, plus Fairlight for audio and Fusion for VFX.

- AI Transcription/Editing: Descript 6.0.2. I tested Otter.ai (great for transcription, bad for editing) and Riverside.fm (overkill for my needs). Descript's text-based editing sold me--it lets you edit video by deleting text, like a Google Doc.

- Short-Form: CapCut 11.0. TikTok's native editor is clunky, and Premiere Rush feels outdated. CapCut's AI tools (auto-captions, clip suggestions) are *insanely* fast, plus it's free.

- Hosting: Vimeo Pro. YouTube's great for distribution, but Vimeo's private review links and ad-free playback make client approvals easier.

Struggle of the day: Overthinking. I almost added Canva for thumbnails (but Resolve can export frames), and Audacity for audio (but Fairlight does that). Lesson: Start simple. You can always add tools later.

Day 2: Tue, July 2 - Recording Day: OBS, Encoding Wars, and Mic Drama

Today was all about capturing footage. I'm creating a 20-minute tutorial on "AI Video Editing Hacks"--perfect for long-form YouTube, with bite-sized clips for short platforms.

Setup:

- OBS Studio 30.0.1 (screen capture + webcam overlay).

- Microphone: Blue Yeti Nano (cardioid mode) with a pop filter.

- Lighting: Neewer 18-inch ring light (5600K, 80% brightness) + a softbox for fill.

First, encoding tests. I wanted smooth 1080p/60fps footage without killing my storage. I tested two codecs:

- x264 (CPU encoding): 1080p/60fps, CRF 23. Result: 2.4GB file for 20 minutes. CPU (Intel i7-13700K) hit 78% usage--choppy when screen recording and talking.

- NVENC (GPU encoding, NVIDIA RTX 4070): 1080p/60fps, CQP 22. Result: 3.1GB file (larger, but manageable), GPU usage 45%, CPU <20%. *Way* smoother.

Winner: NVENC. File size is bigger, but I'd rather have smooth recording than save 700MB. Pro tip: Always test encoding settings with a 5-minute sample before a full shoot.

Audio fail: I forgot to turn off my AC, and the Yeti picked up a low hum. I re-recorded the intro (15 minutes wasted). Lesson: Check background noise *before* hitting record.

Final metrics: 20 minutes of footage, 3.1GB file, recorded in 2.5 hours (including retakes).

Day 3: Wed, July 3 - DaVinci Resolve: Color Grading, Multicam, and Audio Battles

Today's mission: Turn raw footage into a polished long-form video. DaVinci Resolve 19.1 was my battlefield.

First task: Multicam sync. I had 3 sources: screen capture, webcam, and a secondary camera for B-roll (my phone, propped on a tripod). Resolve's "Auto Sync" feature (under the Media tab) matched audio waveforms--*mostly*. The phone's mic was slightly out of sync, so I manually adjusted by 0.08 seconds. Took 15 minutes.

Color grading: The raw footage looked flat. I applied a "Cinematic Flat" LUT (from Resolve's built-in library), then tweaked:

- Contrast +10

- Highlights -15

- Shadows +20

- Saturation +5 (to avoid looking oversaturated)

Audio cleanup (Fairlight tab): The AC hum was still there. I used Resolve's "Noise Reduction" tool (sample noise from 10 seconds of silence, threshold -18dB). Then:

- EQ: Boost 2kHz (for clarity), cut 200Hz (to reduce muddiness).

- Compressor: Ratio 2:1, threshold -12dB (evened out my voice).

Time breakdown:

- Multicam sync: 15 mins

- Rough cut (trimming pauses, mistakes): 1 hour

- Color grading: 45 mins

- Audio cleanup: 1 hour

- Final cuts and transitions: 1.5 hours

Total time: 4.5 hours. Output: 20-minute video, 4.2GB (H.265, 1080p/60fps). It looked *way* better than raw--color popped, audio was crisp. On to AI editing!

Day 4: Thu, July 4 - Descript: AI Magic (and a Premiere detour)

Descript 6.0.2 was next. I imported the Resolve export, and within 2 minutes, it generated a transcript (98% accuracy--missed "DaVinci" once, called it "Davinci").

Filler word removal: Descript flagged 47 "ums," "uhs," and "likes." One click to remove them--*poof*, 3 minutes of dead air gone. Manual removal would've taken 20+ minutes.

Text-based editing: I wanted to cut a 2-minute tangent about OBS plugins. In Descript, I just deleted the text block--and the video cut with it. No scrubbing, no timeline dragging. *Game-changer.*

Studio Sound: Descript's AI audio enhancement. I toggled it on, and suddenly my voice sounded like it was recorded in a studio. It reduced background noise, added warmth, and evened levels. I compared the before/after: night and day.

Export to Premiere Pro: Wait, why Premiere? Descript's export options are limited (only H.264). I needed ProRes for final color tweaks. So I exported a 17-minute rough cut from Descript (H.264, 1080p) and imported it into Premiere Pro 24.3 for:

- Adding lower-thirds (Descript's text tool is basic).

- A final audio pass (limiter to -1dB).

Time saved with AI: ~1.5 hours (filler words + text editing). Total time today: 3 hours. Output: 17-minute "final" long-form video, 5.8GB (ProRes 422 HQ).

Day 5: Fri, July 5 - CapCut: Batch Short-Form Magic

Now for the short-form adaptations. Goal: 5 clips (1-2 minutes each) for TikTok, Reels, and YouTube Shorts, plus 1 LinkedIn snippet (30 seconds, professional tone).

CapCut 11.0 to the rescue. I imported the 17-minute Premiere export and used CapCut's "AI Clips" feature. It analyzed the video and suggested 10 highlights (e.g., "OBS Encoding Tips," "Descript Text Editing Demo"). I picked 5.

Resizing: One click to "9:16 Vertical" (for TikTok/Reels/Shorts). CapCut auto-cropped the frame to keep my face centered--no manual adjusting. For LinkedIn, I resized to 1:1 (square) and zoomed in on the screen capture.

Auto-captions: CapCut generated captions in 30 seconds (95% accuracy). I fixed a few typos ("NVENC" → "NV ENC" initially) and styled them: white text, black outline, bottom-center.

Batch processing: I added a 3-second intro (my logo) and 2-second outro (call-to-action: "Follow for more tips!") to all 5 clips. CapCut's "Copy Effects" let me apply these to all videos at once.

Time breakdown:

- AI clip selection: 10 mins

- Resizing + cropping: 15 mins

- Captions + styling: 25 mins

- Batch effects: 10 mins

Total time: 2 hours. Output: 5 short-form videos (1.2-1.8GB each, H.264, 1080x1920). LinkedIn got a 30-second cutdown of the "AI Workflow Overview" clip.

Day 6: Sat, July 6 - Hosting & Publishing: Vimeo, YouTube, and LinkedIn

Today was all about getting content live--with analytics tracking.

Vimeo Pro: First, I uploaded the 17-minute long-form video to Vimeo (private link, password-protected) for client review. Vimeo's upload speed was solid: 5.8GB file took 22 minutes (100Mbps upload). I added time-stamped comments for feedback: "05:12 - Can we rephrase this line?"

YouTube Studio: Uploaded the same 17-minute video. SEO optimization:

- Title: "AI Video Editing Workflow 2026: OBS → Resolve → Descript → CapCut (Step-by-Step)"

- Description: 300 words with keywords ("AI video editing," "multi-platform workflow") and timestamps.

- Tags: ["AI video editing", "video workflow 2026", "DaVinci Resolve tips"]

- Thumbnail A/B test: Two options--one with text overlay ("18 HOURS TO 5 VIDEOS") and one with a screenshot of the workflow diagram. YouTube's "Thumbnail Test" tool will show which performs better in 48 hours.

LinkedIn: Posted the 30-second snippet with a professional caption: "Building a multi-platform video workflow? Here's how I used AI tools to turn 20 minutes of raw footage into 6 optimized videos (YouTube + TikTok + LinkedIn) in 18 hours. Swipe for the tool breakdown 👇"

Analytics setup:

- YouTube: Enabled "Advanced Analytics" (watch time, audience retention).

- TikTok/Reels: Turned on "Creator Studio" tracking (views, shares, completion rate).

- LinkedIn: Checked "Post Analytics" (impressions, engagement rate).

Total time: 3 hours (uploading + SEO + scheduling).

Day 7: Sun, July 7 - Review, Metrics, and Verdict

Time to crunch numbers and see what worked (and what didn't).

Total time invested: 18 hours (Day 1: 3h, Day 2: 2.5h, Day 3: 4.5h, Day 4: 3h, Day 5: 2h, Day 6: 3h). For 6 videos (1 long + 5 short), that's 3 hours per video--*way* better than my old workflow (5+ hours per video).

Cost analysis:

- OBS Studio: Free

- DaVinci Resolve: Free

- Descript: $12/month (Creator plan)

- CapCut: Free

- Vimeo Pro: $20/month

- Total monthly cost: $32.

Analytics (first 24 hours):

- YouTube: 287 views, 62% watch time (strong for a 17-minute video). Thumbnail B (workflow diagram) had 8.2% CTR vs Thumbnail A's 5.1%.

- TikTok: 1.2k views, 42% completion rate, 120 likes. The "OBS Encoding Tips" clip performed best.

- LinkedIn: 450 impressions, 32 engagements (7% engagement rate--above my average of 4%).

What worked:

- Descript's text-based editing and Studio Sound saved *so* much time.

- CapCut's batch processing made short-form adaptation painless.

- NVENC encoding in OBS kept recording smooth.

What didn't:

- DaVinci Resolve's multicam sync was finicky with my phone footage (manual adjustment needed).

- Descript's export options are limited (no ProRes, forced me to use Premiere).

- Vimeo's upload time was slower than YouTube (22 mins vs 18 mins for the same file).

Verdict: The Toolchain That Stuck

After a week of testing, this workflow is a keeper--especially for creators juggling multiple platforms. Here's my final recommendation:

- Start with OBS for recording (free, powerful, NVENC encoding is a must).

- Edit long-form in DaVinci Resolve (free, Fairlight audio tools are underrated).

- Use Descript for AI transcription, filler word removal, and text-based rough cuts (worth the $12/month).

- Adapt short-form with CapCut (free, AI clips and batch processing save hours).

- Host with Vimeo Pro for client reviews, then publish to YouTube/LinkedIn/TikTok.

Total time per project? ~18 hours for 6 videos. Could I optimize further? Maybe--next time I'll pre-record B-roll to speed up Resolve editing. But for now, this workflow turns chaos into consistency.

*What tools are you using for multi-platform video? Drop a comment--I'm always hunting for workflow hacks!*

-- Alex Rivera, Video Workflow Engineer & Content Creator