Activity
Mon
Wed
Fri
Sun
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More
17 contributions to AI INSIDERS 👨🏻‍💻
why prompt chaining beats single prompts when generating client video assets
generating ai video assets for commercial clients usually breaks down when people expect a single prompt to deliver perfect consistency. the model might give you great lighting on scene one but the character appearance or camera motion completely changes on scene two. here is the 3 step pipeline we use to keep visual continuity across ai generated video sequences: 1. lock style references before motion never jump straight into text to video. start by generating your hero keyframes with identical color grading and lighting parameters. once the visual identity is established pass that exact frame as an image to video input. 2. isolate camera direction from action when prompting motion keep subject action minimal and focus the prompt on camera movement like slow push in or subtle pan. letting the ai handle complex choreography and camera dynamics at the same time is what causes unwanted warping. 3. audio sync and pacing in post ai generated clips lack natural human pacing. bringing the raw clips into a traditional timeline to cut on natural beats and layer organic foley sound is what turns an ai clip into a professional deliverable. what tools are you using for your video workflows right now? curious how everyone handles character consistency.
0 likes • 12d
@Amanda Brown yes i do but not all time
0 likes • 12d
Can you DM me
He had Clay, Instantly, Apollo, and 8,000 verified contacts. He was still at zero clients.
You have the tools. Clay. Instantly. Apollo. A verified list. And you're still at zero clients. That was the guy I talked to this week. Everything inside this community, he had. More YouTube tabs open than I could count. I asked him one question. When you sit down to send that first campaign... what's going through your head? He went quiet. Then: Honestly? I'm scared it won't work. There it is. He didn't have a tool problem. He had an identity problem. Most people in this community are sitting on a loaded weapon they're afraid to fire. Not because they don't know how. Because somewhere between the masterclass and opening Instantly, they decided the gap between them and a working campaign was knowledge. It's not. The gap is a decision. The people already eating aren't the ones with more tools. They're the ones who decided they were already the kind of person who sends the campaign. I built $1K MRR off a database that got called garbage. Not because I had better data. Because I stopped asking whether it would work and started asking what I'd learn if it didn't. That shift is worth more than every Clay credit you'll ever buy. There is a version of you that already has everything you need. It is just waiting on a decision. 👇 What do you sell, and who buys it?
1 like • 15d
tools dont close clients proof of work and relevant problem solving do. sending 50 tailored loom video teardowns of a prospects existing content or ads outperforms 5000 generic clay sequences every time. when you show them their exact bottleneck fixed on screen the fear of outreach disappears.
0 likes • 14d
around 15 to 20 highly focused ones per week. instead of spraying volume we pick creators or brand channels where their hook drops off in the first 5 seconds or where caption contrast hurts retention. we record a 60 second screencast showing the exact timeline fix and drop a link to a recreated 10 second sample. response rates hover around 25 to 30 percent because it proves capability before asking for anything.
I let Claude Code run 1,000 inbox placement tests
Every deliverability theory I'd picked up over the years, I tested. Claude Code plus an inbox placement API, testing variations and combinations, on Google and Microsoft. Over a thousand tests later, I had a rule set and a Claude Code skill instead of a pile of opinions. The point is the method, not my rules. Your emails have your own edge cases: Loom links, GIFs, images, whatever you send. My rules are a starting point. Lots of placement tests (essentially unlimited on the paid plan) plus AI writing the variations is how you find yours. One result I still can't get over: a single change to how a link was written moved all 11 test emails from Promotions to Primary. The video shows what it was, plus the Microsoft habit most placement tests miss. Full video: https://www.youtube.com/watch?v=ByrOheLnWUY&list=PL0Yyhj9Mcbg8zog1pSTOTVfmyHKbtRClF Go watch it, then tell me what you'd have Claude test first.
0 likes • 15d
systematic split testing over guesswork always wins. embedding loom links with animated preview gifs usually gets caught in promotional tabs if dkim and spf headers are not calibrated for rich media. testing placement with ai scripts before blasting cold campaigns saves the domain reputation completely.
🤖 Let's Talk About AI Tools!
I want to know which AI tools are best for different types of work. Here are 6 popular AI tools: 1. ChatGPT 2. Claude 3. DeepSeek 4. Google Gemini 5. Perplexity AI 6. Microsoft Copilot I'd love to know from your experience: - Which tool is best for coding, writing, research, design, video editing, data analysis, automation, etc.? - How can we get the best results from each tool? - Which plugins, extensions, or integrations can make these tools more powerful? - Which tools offer the best features for free or at a very low cost? Also, if you know any other useful and almost-free AI tools apart from these six, please share their names and explain what type of work they are best for. Your recommendations and real experiences are welcome! 🚀
0 likes • 15d
ffmpeg paired with whisper timestamping has been our most reliable backbone for client video workflows. shiny ui tools come and go every month but raw command line pipelines give you predictable rendering speed and zero monthly seat costs when batching ad creative
Why your Claude output is weak (and it's not the model).
Most people prompt, then complain about the output. The fix is one sentence, and it goes in before anything gets built. When I had Claude build a full high ticket funnel, I gave it the broad instruction first. The offer, the price, the pages I wanted, the sequences behind them. Then instead of hitting go, I put it in plan mode and added this: Interview me to clarify any major points, but use recommendations for most. That second half matters. Claude makes the small calls itself and only comes to me on the major ones. Here's some of what it asked me: What should this $20K program help clients achieve in 90 days? Which delivery model should the funnel sell? Who should qualify to apply for this program? How aggressive should the core 90-day promise be? Those are all offer questions. It was making me think through my own offer before it built a single page. If you're not sure what questions you should be asking about your offer, that interview is the exercise. I rushed mine because it was a demo. On a real build, the more time you spend in the interview, the happier you'll be with the final result. Drop your biggest Claude prompting mistake in the comments. What did you skip that you wish you hadn't?
0 likes • 16d
treating the context window like an editing timeline where you strip out unrendered clutter before running the next prompt completely changes output fidelity. passing tight structured json schemas instead of open ended briefs makes the execution reliable every single time
1-10 of 17
Saad- Khan
2
7 points to level up
@saad-khan-4361
Business Person (AI Video Editor Creator) at Oladoc Lahore From SKP | Future CEO | Building eCommerce Agency | Always Learning | DM for Collaboration

Active 10h ago
Joined Apr 16, 2026
Lahore
Powered by