Creating Videos with AI in 2025
A Hands-On Look at the Current State of AI-Enhanced Production
What started as a casual favor for a colleague turned into a deep dive into today's AI-powered video tools. I produced multiple short videos, combining scripts from ChatGPT, voice-overs from Artlist.io, stock footage from Envato, and editing in Final Cut Pro. The process was surprisingly fast, impressively capable, and full of lessons about what AI can (and can't) do in creative work.
August 2025
In mid-July, I found myself diving head-first into the world of AI-enhanced video production, as a favor and out of sheer curiosity. What started with a casual MS Teams message quickly turned into a small production sprint and, in the process, became an unexpected journey into the current state of what AI can (and can't) do for creators today.
For context: I'm a web developer. I don't work in video production, and I don't moonlight as a filmmaker. I just find that video is an increasingly important medium for communication, and I wanted to see how AI could help streamline the process. A lot has happened, since I last dabbled in AI video production in 2023. I absolutely love that producing videos that look professional is more accessible than ever (even for someone like me).
A colleague from our company's AI evangelism team posted on our shared channel looking for someone who could produce a short video for the sales department. The goal: to show how AI could help improve outreach and analyze sales data.
I replied that I'd be happy to help. A few minutes later, she sent over a script for the voice-over narration and some rough ideas for visual direction. And just like that, the creative process began.
I used Artlist.io to generate the voice-over narration using
the AI text-to-speech feature. The results are uncanny!
Building A Video: A Hybrid Workflow
I needed video footage. Fortunately, I already have a subscription to Envato Elements. For me this is one of the best-value tools for creators thanks to its massive library of stock footage, sound effects, music, and even 3D assets. I had a solid foundation to start.
However, as good as stock libraries are, one of their core limitations quickly became clear: continuity. You can't follow a specific character across scenes because each clip uses different actors, lighting setups, and locations. That's fine for general themes, but if you're trying to tell a linear story with a single protagonist, it becomes tricky fast.
Unfortunately, Envato's AI text-to-speech left a lot to be desired. So, I turned to Artlist.io, where I have a paid subscription. I made minor edits to the narration script and pasted it in. Within minutes, I had a voice-over that sounded spookily natural... smooth, believable, and consistent over a full two-minute stretch. It's wild how far this tech has come.
That said, text-to-speech has its own limitations too. You can suggest emotional tone or pacing, but what you won't get is a performance. There's no AI version of Meryl Streep or Nicolas Cage delivering nuance and unpredictability. What you do get is "good enough" for informative content. But it isn't going to carry a drama or deeply persuasive pitch (at least yet).
I doubt that AI will replace the unique qualities of Nicolas
Cages performances any time soon.
With the narration in hand, I pulled together suitable stock footage clips from Envato and Storyblocks, and sourced fitting background music. After a few disappointing attempts with Google Gemini-generated footage, I settled on using just three short AI-generated video segments.
The intro and outro came from Viddyoze.com, a web-based tool I paid for over seven years ago. To my surprise, my lifetime license still worked, which felt like a small miracle in our current age of endless subscriptions.
I created custom titles in ArtText4, and then stitched everything together in Final Cut Pro on my Mac. The end result: a clean, effective two-and-a-half-minute video that communicated its message and looked professional.
This is the timeline in Final Cut Pro, showing how I arranged
the clips, titles, and voice-over.
Creating a Personal Take on Cloud vs On-Premise
Since I couldn't share the internal video publicly, I decided to create a similar one to test the process again and document it in this article. This time with a topic I care deeply about.
Even though I work on a cloud-heavy product as a web developer, I believe the cloud isn't always the best solution. Many companies can benefit from hosting their own services and data on-premise. It offers more control, flexibility, independence, and lower costs (under the right circumstances).
So I opened ChatGPT and quickly outlined a script based on that premise. Again, I used Artlist.io for voice-over (trying a different AI voice), designed titles in ArtText4, pulled stock footage from Envato, and assembled everything in Final Cut Pro.
It took about four hours to produce this particular video end-to-end. No, it's not cinematic art but I think it is polished, clear, and delivers the message. It does the job.
I pulled stock footage from Envato for the video's visuals.
A Different Use Case: Summarizing Product Demos with AI
Not long after, another opportunity to use AI in video editing came up. One of my managers asked me to edit two demo videos into a single coherent walkthrough. These were screen captures narrated by a colleague, explaining how a new application works.
I hadn't seen the app before, so I needed a way to understand and introduce it efficiently. While watching the videos, I recorded the audio on my company iPhone using the Voice Memos app, which automatically transcribed the speech. I transferred the transcript to my MacBook Pro and used Apple's built-in AI tools to generate two summary paragraphs: one for each section of the video. After minor edits, those summaries became the on-screen intros.
I combined everything in Final Cut Pro: title card, videos, intro text, and end card. The final product was simple but professional, and it came together quickly thanks to AI-assisted summarization.
I designed the titles for the video using ArtText4.
Advantages and Limitations of AI in Video Production
These projects showed me that AI isn't just hype when it comes to video creation, it's already changing the game in very tangible ways. From writing scripts and generating voice-overs to summarizing content and streamlining editing, AI tools can dramatically reduce production time and effort.
But more than that, they encourage iteration. I don't get overly attached to a version because making another one doesn't feel like a major time investment. That's huge when you're working under deadlines or exploring multiple creative angles.
There are still limitations, of course, especially in areas like AI-generated video footage or nuanced voice performance. Overall, the state of AI in video production is impressively usable already. And it's only going to get better.
From Concept to Creation
What used to be a full production process requiring a team of people can now be done by a single person with the right tools, most of them AI-powered.
It's not just that AI makes things faster. It changes how I think about creativity, effort, and scale. Whether I'm producing internal explainers, promotional material, or conceptual storytelling, AI can be an incredibly powerful ally.
I'm excited to see what comes next.