Complete newbie's first video - it's bad. How do I improve quickly?

Your first video is like the first span in a poorly instrumented trace, you get the basic structure but the detail is missing. Here's where I'd start.

On the voice, always record your own. Tweaking AI settings is a fool's errand, you'll spend hours chasing a natural cadence you can achieve in three takes with a decent USB mic. The robotic tone is a symptom, the cause is a script written to be read, not spoken. Write it, then read it aloud and slash every other sentence. If it sounds formal on the air, it's wrong.

For visuals, free assets are fine but you're using them wrong. A boring slide is just a span without tags. Don't just show text, show the thing. Screen recording a graph, a diagram from your docs, even a terminal with a command, it's all more engaging. Pacing is tied to this, you cut away from your face or a static slide to these visuals, which gives the viewer a break and hides your edits. The script should be a series of beats, not a monologue, each beat paired with a visual change. If you're talking for more than 15 seconds without a visual shift, you've lost them.

Spoken word scripts are indeed the root cause, but the "slash every other sentence" advice is a bit romantic. For a newbie, that just leads to incoherent rambling. You need to replace the formal phrasing, not delete it. Swap "we will now proceed to examine" with "let's look at". It's a vocabulary transplant, not an amputation.

And while showing the thing is good, telling a newbie to screen record a terminal or a graph is skipping five steps. First, they need to learn how to *briefly* show that thing. Otherwise you get a three-minute static shot of a CLI while they stammer. The visual shift isn't a magic trick, it's a timing discipline you have to build first.

Oh gosh, I feel you on this. My first explainer video for my little bookkeeping service sounded like a bored airport announcement. I think your instinct about pacing is the real key, more than the voice or visuals even. If the pacing feels weird, nothing else works.

For the script, I found reading it into a voice memo app first, just on my phone, helped me find the rushed parts. You hear yourself tripping over certain phrases immediately, and those are the ones to rewrite or break up. It's okay to sound like a person, not a news anchor.

A quick win I stumbled on for visuals was just using more zooming in and out on Fliki itself, even on simple slides. It adds a tiny bit of motion that keeps things from feeling static, and it's built right in. Makes free assets feel a little less flat while you're learning the rest.

I think you're right to break it down into those three categories. It makes the problem feel less overwhelming. Since you're asking for top priorities, I'd actually suggest you tackle them in this order: pacing, voice, visuals.

Why? Because a rushed or slow script makes even a great voiceover feel off, and it dictates what visuals you need. Record yourself reading your script on your phone, then listen back. Anywhere you stumble or take a big breath is a cue to shorten a sentence or add a pause. Write for the ear, not the eye.

On your specific voice question, I agree that recording your own is the fastest path to sounding human, but I understand the hesitation. A quick middle ground is to use the AI voice to generate a first pass, then record your own reading of that cleaned-up script. You'll be more confident, and you can inject your own emphasis. The robotic tone often comes from a flat script, not just the AI.

For visuals, the best quick win with free assets is to introduce simple, consistent motion. Use gentle zooms or pans on a single slide, or transition between two similar images to create a 'build' effect. A static slide for 15 seconds feels dead, but the same slide with a slow zoom over that time creates focus and energy. It's a single setting to learn that changes everything.