Gemini (Nano Banana + VO3) + Gamma + Descript Tutorial | Event Promo Video Workflow

Full video transcript

Hey, how’s it going everybody? Welcome in to today’s Gamma Mastery Workshop where we’re going to be pairing Gamma with Gemini for a powerful combo of the two platforms. See how we can use them together to do all sorts of incredible stuff with AI. Good to see everybody here this fine Tuesday. How’s everyone doing? Good to see you, Mark, Sawen, Mohamad, William Russell. Good to see everybody. Before we kick things off here, I’m curious for those of you here that have been following along with our Gamma series so far.

How has Gamma been going for you since the start of the series? Any breakthroughs? Any wins or any aha moments from the past 3 weeks in getting your Gamma account set up and getting content, presentations, that sort of thing going? Any wins out there before we dive into things today? I’ve been using Gamma since you first introduced it, I think almost a year ago. And I use it for website pages, presentations, decks, slide decks, sales presentations. It’s a fantastic program. It saves me a phenomenal amount of time.

Yeah, it’s a great tool. I’d like to learn how to use it better. And I’ve been using a lot of Gemini lately. Now that it’s 3.0 and we have that Nano Banana, it’s generating a lot of images for me. I even tried some videos. I’d like to know how to string those videos together. They’re only 8-second clips at this point in time, but I’d like to be able to figure out how to put them all into one complete video, those multiple 8-second clips. Well, are you using any video editing software tools right now? No, I’m not.

Right. Cool. We can touch on that if I don’t get to it in the context of today. Just if you want to just drop a note in the chat so I can keep an eye on it. I do have a workflow that I use Descript to put the video clips together and edit it in that front, but there’s probably a whole video just for that we can get into maybe in a future date. Cool. Thanks. I used to use Descript, but since the upgrade now I’m getting a different screen, I’m having a hard time adjusting to the technology.

Maybe if you go over it one day, that would be great. Sure. Yeah, we can maybe we’ll see where the day takes us when we dive into things. Thank you for sharing, Mark. Any other wins in your… I used to use Gamma very primitively — just put in something and generate presentations for my lessons and my teaching — but I figured out how to change the theme, because I used to start over every time I wanted to change something. Now I know how to edit; to me it’s a big win because I created some social presentations to attempt to post on TikTok or something and I didn’t like it and I couldn’t go back to the regular presentation.

But after watching the instructions, to me it’s a big win because I don’t have to go again and create something new. Yeah, that’s great. It’s a great one to highlight as well for anyone that’s jumping into this series right now that hasn’t gone through the foundation of setting up a brand style guide in Gamma — that’s where I’d recommend everybody starting if you’re going to start using Gamma for content: spend a little extra time going through setting up your brand style guide, your own template, with your own brand colors and your own vibe, so that you don’t have to recreate it every single time.

It’s a massive timesaver down the road and doesn’t take too long to get set up. The first video in the series covers that entire process inside of an hour. And we’ll get things kicked off here in just a moment. But any other wins from anyone else here tuning in with us live? Assad, good to see you. Just thought you’d join in. Any other wins before we kick things off, or any burning questions that you’d love to see covered today in terms of Gemini and Gamma? This will be our first of this kind of pairing of Gamma with another tool in the series.

And so I think setting the context for this specific workshop: when we’re walking through the generation of any sort of content, any sort of use inside of Gemini, you’re going to see some parallels to how you can apply the same sort of methodology, the same workflow, the same way you can prompt Gemini, and see how that translates to other tools like ChatGPT or Claude from Anthropic. Now, Google, I think, has been creating a lot of value from the tools that they’ve been releasing over the past year or so.

And with the release of Gemini 3 lately, I’ve been seeing a lot of big-time improvements — but also where Gamma is fitting into this mix, integrating certain pieces of Google’s tools like Nano Banana, for example, so you can generate images with Nano Banana Pro directly inside of Gamma, which is super powerful. And so I want to walk through and just dive into it to kick things off today. Make sure my proper screen is pulled up. How we can go about using some of the core features inside of Gemini that, maybe if you were seeing some of these tools released a year ago, you may have had to go to multiple places — like a V3 platform — to generate videos or images, things of that nature.

But one thing I was generating right before we hopped on this call is a cool example of how you can use Gemini when you’re inside the Gemini workspace. It looks very similar to how you’d be set up with a ChatGPT workspace. You have your chat screen right here in the middle. You have an expanding menu where you can access your different threads. And you can even create these things called Gems, which I haven’t gone super deep into, but my understanding is Gems are Google Gemini’s version of what a custom GPT is inside of ChatGPT.

So, expect us to go a little bit deeper into creating Gems inside of this workshop series. But when you’re creating inside of Gemini specifically, there’s a few prompts that can get you started that I think are just good to understand what this tool can do, because it can seem like a blank canvas — it can do anything. What can I even get started with creating? One, it can create images. If you’ve seen anything about Google’s Nano Banana, it is a powerful AI video generation tool that can also edit images.

And I found it to be particularly useful in editing an image while keeping the rest of the image intact and then adding new things into an image or removing certain things from an image. But also, obviously, you can write, create, brainstorm, research — which Google is a great tool for doing, things like deep research — but you can also create videos directly inside of Gemini. And so if you’re curious how you can use Gemini to create a simple 3 to 8-second video, before we hopped on this call…

I just took a quick thumbnail example. We’re building out some thumbnails for my Side Hustle game company, which I downloaded from Gamma. I just took a thumbnail and said, “Hey Gemini, make a 3-second video from this Side Hustle thumbnail, animated.” It took a minute or two and then it came out with this. Can you guys hear that, the sound that I just played? Okay, I’m going to see if I can share it with my sound, because that’s one of the cool things I found with VO3 and part of Google’s AI suite — that again you can access it now directly inside of Gemini, which is why I’m showing this. How do I show my screen share sound?

It actually creates sound with it too. And so just to give you some inspiration for what’s possible: this was an image that I took from Unsplash for the background — so a free image that you can get from Unsplash, literally millions of images on there. I just added my logo and said, “create an intro.” So cool — now I can have a quick little intro. I could probably prompt it with more of the type of sound that I wanted and play around with this a little bit more.

But now if I said, “okay, that’s cool, Google — let me select create images.” Now I started with the seed image and said, “Can you create another image for me based on the image I provided that is just optimized for a 300x300 image style, but keep everything else the exact same?” So what I’ve done here is I’ve given it the initial image for context, had it create a video based on that image, and now I’m just asking it to keep everything else in this image intact but resize it for me in a specific size and style.

So, did it do it? Doesn’t look like it did it exactly the way I wanted. So, a little bit of a thumbs down for Gemini in this case — it didn’t quite get to where I wanted it to go on that front, but maybe it was because I didn’t give it clear enough instructions, or maybe I could have prompted it differently. So I’m going to show another way I’d approach this if I wanted to create additional content based on this thread and this context inside of Gemini.

Let’s say I’m going to prompt it again: “I’m going to create an event for New Year’s Eve that’s going to be a private party where we’re going to play the Side Hustle game and raise a bunch of money for a charity event. Can you create an image that goes along with this event and then a description of this image that we can eventually turn into a presentation? I want to have multiple slides that talk about the event and make it super exciting for people to join.” Going back to how we as humans are operating LLMs and interacting with these tools in a conversational way —

this is a very similar approach to how I might approach creating something with Claude or ChatGPT. But I know that I’m using Nano Banana here, and Nano Banana is really good at creating images directly inside these chats. And one thing I found is that the images — the words, the styles, that sort of thing — compared to a year or two ago: you guys remember when AI images first came out, it was like writing the word “boring” and it would come out as alien speak, misspelled letters everywhere.

Well, nowadays it’s pretty on point with creating text in images at that size. So while I kept it in the same style of the image, you can see: New Year’s Eve, Side Hustle Smash, raise a glass, raise some cash, play for a cause. I’m just coming up with this on the fly right now. I can see it’s still on that thumbnail style, which isn’t a problem, but I can see “raising funds for [charity name].” And now, for editing images inside of Nano Banana — because I’ve done this with other images I’ve created in Canva or generated with other models and used Nano Banana to optimize —

you could say, “this looks really good on the right-hand side where you have the charity name section on the screen. Can you change the charity name to ‘Hustle for Kids’” or something like that — “raise your money for hustle for a cause.” Let’s say “Hustle for a Cause.” And so how this should work, based on how it’s worked for me in the past, is: keep everything else in this image the same, but just update that one particular section. So now we can look at these top to bottom, and they’re pretty much the exact same image.

Nothing else was really changed at all in either image, except “Hustle for a Cause.” So not perfect, but you can see how specific it was with editing that one section of text. So again, conversationally providing prompts, editing the image — something you can do directly inside of Gemini as a thread. But now, let’s say, “okay, this is good enough, but I actually just want you to optimize those words to only say ‘Hustle for a Cause’ to make sure it says exactly that text.”

“And then produce that image for me and we’ll create the rest of the content for the marketing material.” So it’s loading Nano Banana again. And my bet is that once I have an image right here, I can just jump right back into VO3 and say, “Hey, animate this image — have everyone raise their glasses up and present a toast.” That could be the introduction slide for a video. It’s missing the mark right now on that front, which — it’s all AI, generative AI can make mistakes. Double check it.

Make sure it’s actually producing what you want. But we’re going to take this and keep rolling with it. We’re going to say, “create a video for this image.” Let’s see — is it going to let me do that? “Start a new chat” — no, I don’t want to start a new chat. That’s weird. “Can you use VO3 now to create an animated version of this where everybody at the table is raising their glasses up?” Interestingly, whenever I went to select the actual tool I wanted to use, it was trying to force me into a new chat.

But when I just spoke it — said what I wanted it to do — it went through the process. Some of this stuff is totally new to me; this is the first time I’m trying this particular workflow inside of Gemini, swapping between these two tools, because I didn’t realize I could use VO3 directly inside of Gemini until just before this call. Key takeaway for anyone watching: these tools change constantly. How powerful Nano Banana is today is going to change in several weeks — how good the image generation is, how good the video generation is, that sort of thing.

It’s going to take a minute or two to create this type of video. But what was the question we had earlier — was it Mark, about using these types of clips and stringing them together to create other videos from it? You can create these 8-second clips, but how do you take those and put them together into a fully fleshed-out video? A quick question came in: if you take a stock photo that has copyright and ask AI to make a change to it, once it’s been changed but more or less remains the same, is it still covered by copyright?

It’s a great question that I’ll preface by saying I’m not a lawyer and don’t play one on the internet. I’m not entirely sure, but I’d say if you’re going to use images and modify them and then create your own stuff with them, it would make sense to check the copyright of those images and maybe consult a lawyer if you’re going to do anything significant with them. The reason I mentioned Unsplash is because those are all royalty-free, completely free images. I’ll have to check with a lawyer on the exact rules around modified stock images.

I think there are probably some workarounds — like taking reference images and building your own AI versions based on those — but it’s worth checking into if it’s a concern. Let me see if you guys can still hear my video. This is now just an 8-second clip, with a different intro sound. It actually skipped the image I wanted it to create the video on and went back to the original one. So what I think is happening here is that Gemini was just referencing this image for the main thing and not really listening to what I wanted it to do.

So, for one last example before we move on to creating a presentation in Gamma based on this, I’m going to rerun it through this prompted context and then walk through what I’m going to do as part of the next steps in the creative process for a full presentation. I don’t want to spend a bunch of time prompting and reprompting beyond this last one — I do want to see what it can do. Some of this is trial and error, testing to see how we can push AI to do what we want.

Maybe it didn’t get the hint, but taking the concept and breaking it down: we have an introduction slide, we’ve got the concept for this New Year’s Eve party, and then we need to create a few slides talking through what people are going to experience at this event — what they’re going to get out of it. This event could be anything in your own world with how you’re going to use these tools — a private workshop, a class you’re teaching, a training presentation. Think through your exact use cases and what’s most relevant to you.

But the fundamentals of working with a tool like Gemini as a large language model — creating images that serve as reference points to build additional slides or prompts from, and creating outlines and creative briefs that we then take into a tool like Gamma — will directly influence the quality of the content Gamma ultimately produces for us.

And then, if we have time today, a very quick workflow for how you could take that content and a series of slides presented in Gamma and move that into a tool like Descript, where you can start to edit and clip these slides together. You could also use other video editing tools to get that process going. I see a lot of these pieces as sections inside a workflow: I want to create a 30-second trailer that totally shocks and stuns my audience and makes them want to learn more and sign up for this event.

That’s going to cost you $10,000, or it’s going to cost you two or three weeks of lead time and prep to do all these things another way — or we can take another version of that and use AI as the way we’re leveraging things. Okay, this one’s done — looks like it did it the right way. Let me readjust my screen and share the sound. “Happy New Year, everyone. Cheers.” All right, cool — so now we’re ready for a real Side Hustle New Year charity bash.

So cool. Now I’m feeling a little more excited about where we can go here — I’m starting to see some of this direction. Don’t forget who’s operating these AI tools: it’s you. It’s the operator, the human behind the prompts, behind the desk, behind the screen, actually instructing and directing where we want to go next. So where do I want this process to end? What’s the vision I have that I want these tools to help bring to life? I’m going to talk through exactly how I’ve been doing this to create these sort of presentations with Gamma from inside of LLMs.

This is the first time I’m doing this process here inside of Gemini specifically — I’ve been using Claude heavily for this before. So I’ll be interested to see how it produces things here. By the way, I’m using a tool called Whisper Flow, which I use all the time nowadays, for voice dictation. It’s super effective at getting words out and actually capturing them without a bunch of misspellings. So I’m going to say, “Okay, this is great.

Now, what we want to do is create the creative brief for a presentation that we’re going to put together telling people about this New Year’s Eve Side Hustle Smash event. It’s going to be a charity fundraiser, have some fun, raise some glasses, raise some cash — raising funds for a cause that will benefit youth and entrepreneurship in the city of Austin. We’re going to be inviting a lot of entrepreneurs, founders, and investors to have fun and hopefully raise a lot of money for this cause.“

“And it’s going to happen on New Year’s Eve this year, 2025, which is December 31st. It’s going to go from 6 until past midnight. What we want to do is create a 10-page slide deck outline that talks about who this is for, what we’re going to be doing, why this cause matters, that tickets are invitation only, and that you’re getting this presentation because we wanted to invite you to the event — ending with a QR code they can scan to purchase their ticket. Talk about the dress code.”

“And inside these prompts, for each slide, I want you to create a complementary description for what that slide is going to be, locking it into the entire theme of what this event is going to be like. And from the very start, we need a prompt for a custom art style guide that keeps the theme and custom style in line with this image and presentation — classy, fun, going to be a blast, going to be a great event.”

“So I need you to create all that content right now — go ahead and execute it.” So I’m giving it my prompt — it took all of that from a voice prompt I was thinking through: networking, fun, philanthropy, New Year’s Eve — “create a brief, 10-slide deck outline.” Just reiterating back what I gave it so I know it’s tracking with what we want to create.

And then it’s going to create the presentation style guide. The reason I’m starting here and then pairing this with Gamma is because inside Gamma, when we get to the slide deck presentation, you can give it a way to produce custom art in a style guide. I have access to Gamma Studio right now, which some of you may or may not have depending on your plan, but it uses these types of style guides and prompts and images in a more informed way to create the types of images we’re looking for.

So we have the 10-slide deck outline: slide two, why you’re here; slide three, when and where; slide four, main event, the cause — Hustle for a Cause, entrepreneurs, investors, the vibe slides, call to action, thank you, and contact us. “Can you give me a specific custom AI style guide art prompt that will match the vibe to keep this consistent across every single slide? I want this locked in before I start.” It’s loading Nano Banana AI art style guide prompt.

I wonder if it took this instruction as a way to create an image for it — if so, I’m just going to say “no, I just want the words.” So it built out a bunch of slides with this, which by itself isn’t a bad way to go, just not as dialed in as I’d like. This is AI being AI — it may be assuming a little too much, thinking it should just go and do this for me.

And that’s not what I wanted at all — I need to dial it back a little bit. So it’s not being too helpful right now, Gemini, but we’re not going to let that slow us down — we’ve got a lot of what we’re doing already built out. At the end of the day, this is what I’m looking for: we went through — recap — giving a seed image, city of Austin, here’s the brand logo, here’s the vibe.

“Here’s — I want to create an event, right? Give me an image based on this event.” So we did that — we have an intro now, and a little teaser: “Happy New Year. VO3 is actually including sound.” Now we have a full creative brief with an explanation of every single slide we’re going to be creating. There are two ways I could go about demonstrating this inside of Gamma, taking what we have here and creating a full presentation strictly based off of it.

But we’re also going to be using Nano Banana to produce these images — so we’re going to Gamma, but still using Google Gemini Nano Banana Pro inside of Gamma. I’m going to tab over to my Gamma account, go to my Side Hustle folder, and click create new. From here, the two paths I’m going to share are: pasting in text as a way to start creating one of these, and then demonstrating Studio, which — once I do that — you’ll see the difference between what Studio produces versus the regular paste-and-text flow.

If I go to paste-and-text, I’ll select presentation first, since that’s what I want to create. Then I’ll come back to this 10-slide deck creation, copy all of it, and paste it directly into Gamma, saying “preserve this exact text” and configuring my prompt. Now Gamma takes the text I gave it and moves it into the prompt editor — and it’s also split the cards up, which is handy, because now it’s card by card, each one an individual card.

I can either preserve it or generate it — I’m going to select generate in this case, because you can experiment: test it, run it, bring it back, run it again to see how it does. This itself is a much more structured starter prompt that we spent time putting together with Gemini, which now feeds Gamma’s AI as well. I don’t think I need super detailed amounts of text here.

So I’m going to make it far more concise. You can see it highlighted “who am I writing this for” — influential Austin-based founders, investors, entrepreneurs invited to this event, energetic, polished, celebratory. That’s cool — and now I’ve got my theme. Taking it back to the beginning of this call, we talked about themes and starting your brand style guide — I don’t need to recreate this theme or start from scratch; I’ve already got my Side Hustle theme locked in.

I’m selecting that theme, then choosing image sources. Then you have Nano Banana Pro — there’s a big difference here. Nano Banana Pro just recently got released inside of Gamma Suite, and you really need to pay attention to the credits per image. It’s showing 10 credits per image right now, slashed down from 70 credits, which — depending on what plan you’re on in Gamma (if you get 4,000 credits a month) — every image could be anywhere from 10 to 70 credits depending on when you’re using this.

This may not always be slashed to this price, but they’re rolling it out and want people to test it, so we’re using the Pro version. Be aware, because it can be very easy to burn through a ton of credits — if you’re generating a lot of these and you run out, it may be because you selected a model that used significantly more credits, as opposed to other tools that cost around 3 credits per image (for example, image editing at around 2 credits per image).

Credits for these plans regenerate each month — the host mentioned running out of all their credits during a Gamma API contest a week earlier, and having their monthly membership renew and replenish credits over the following weekend. Last thing before generating: with image source set to Nano Banana Pro, I want to add the finishing touches by adding the style guide.

So I’m going to copy this style guide and paste it into the custom art style field, and let’s see what we get. This is going to be the first generation — we’ve got our 10 slides locked in, Nano Banana Pro as the image generator, and we’re using the template we already spent time setting up. Using Gamma’s traditional generator, we’re going to have the starting point of what could be a very solid presentation for inviting people to this event, that you could walk people through in a two- or three-minute Loom video.

The workflow here can be applied to generating any sort of presentation or style — it just depends on how you’re outlining and providing context in advance. Already on image number seven — cool, dress code slide. So we see here it didn’t have my company name, event email, or phone number, so it gave placeholders — but I can easily modify this. We got: cool, Side Hustle New Year’s Eve smash, raise a glass, raise some cash, play for a cause.

We got an intro slide, an invitation slide with all the details — location, secret until the ticket is confirmed. This slide doesn’t quite look like what the game actually looks like when we’re playing it, so I might want to update this image with Nano Banana. There’s definitely more time I’d spend before taking this to a final piece where I’d start inviting people to it — maybe an hour more to get the rest of the images locked in.

Overall, it laid out the context here inside a 10-page presentation pretty much exactly how I wanted it to start. So for me, Gamma did its job very well for getting this first version generated. Now I can edit and modify it, but I didn’t have to spend all the time organizing and putting these slides together myself. If you’re good with this as the final output, you could go in and edit it, modify things, and change colors to get it to a final piece.

You can also go to the upper right-hand corner and use agent mode — working with Gamma as an agent to give it really specific instructions, like “highlight the 2026 in this first slide in our on-brand green color, and also highlight the words fund, finance, and philanthropy in that green color.” And then I can just talk to it to have it do anything else. Reading back a few questions: someone asked if you can export to PowerPoint — yes, you can export to PowerPoint or Google Slides so you can continue editing there.

Just from prompting it, it made those slight color updates. I’d highly encourage getting more comfortable with agent mode in Gamma and playing around with it, especially if you like to use voice dictation for modifying things. This agent is only going to get more powerful and know more about all the other Gamma tools you can see in the side box, and there’s a lot of instructions you can give it to modify and edit slides.

Once you’re ready to export, go to the top navbar, click share, and you can export a PDF, PowerPoint, slides, or PNG. You also have the option to export directly to platforms like LinkedIn. But if we’re going on to the next step of turning this into a video, I would export these as PNGs and upload them into Descript. I want to show that process, and contrast it with generating images and presentations inside the regular Gamma copy-and-paste flow versus doing it inside Gamma Studio.

I’m going to go a little faster-paced here to be conscious of your time and recommend just going back and watching this again. We’re already coming into this with the assumption that you’ve got your outline ready — in this case using Gemini as the starting point. Back in my Side Hustle folder, I’ll click new, but introducing Studio mode, which is in beta — use images to tell a story with immersive visuals and text that flow together.

You can see immediately it looks a little different in Studio mode. I’m limited right now to 20 cards, and I’m going to select 10 since that’s how many we’re generating. I’ll select page default to traditional, and I’m going to add a little context, because I don’t want it to just be straight-up images. I want a little bit of context added in. This is the difference in Studio mode that’s important to highlight.

Going back to my Gemini thread, I’ll give it this prompt again, the same way I started generating it inside Gamma the first way. I’m describing what I want to make, with a little context, and it’s going to automatically generate the outline. I’ve been playing around with just leaving the output at these outlines versus quite literally copying and pasting each individual slide prompt.

One thing I’ve noticed: this context generates the outlines, but I’m still experimenting with the quality difference between leaving these prompts as Gamma generates them versus literally copying and pasting slide title, body text, and visual description for each slide. I’m still testing it out, so I’m going to leave it as-is for the sake of time, but you could alternatively copy in the exact fully detailed prompt.

Side Hustle 4.0 is my visual theme, and it kept the same color palette I had before, along with Nano Banana Pro since that’s what I’ve been using in Studio. This is going to take a little longer — maybe 30 to 45 seconds to start the initial process. Studio is much more visual-heavy, and the workflow to generate this amount of images directly inside a Gemini chat would take significantly more time than doing it here. I’ve been happily surprised by how well the text has been showing up and generating correctly with this format and flow.

Once it generates, you can also edit the images with the same agent mode — it just appears a little differently. Let’s see what Gamma, tapping into Google Gemini Nano Banana Pro, generates. That’s a really strong image — this looks a lot better than what we were trying to generate directly inside Gemini earlier in the call. It might not be perfect, but you can always go back and edit, modify, and refine. The more time you spend on the prompts for each individual slide, the better the flow you’ll find for the images you want.

I might want to remove this Side Hustle image and add my own logo in post-production instead — so I’m going to ask it to remove the Side Hustle logo in the top-right corner. And that’s not there anymore, because I selected this first card, it edited the image directly using Gamma’s agent to help navigate this. That looks shockingly like the Austin skyline — that building is actually there, the Frost Tower.

I could also ask it to remove the card deck graphic in the middle and display a different version of the game, or add that in post-production myself, or download the image and have Gemini update it further. I’m glad we have two other workshops planned going deeper on Gemini, because there’s so much to cover. But hopefully this gives you ideas for how you can use Gemini with Gamma together in a workflow.

Importantly, Gamma is plugging in and making it easier to create really strong content connecting with founders and capital. This has me excited to maybe host an event like this at some point — it really elevates the whole vibe for a fundraiser like this: game night, high-level, exclusive, tons of fun.

Now, to go through the video process, because every good video needs some lightweight background music: I’m pulling up my Suno account and prompting it — “we’re putting on a Side Hustle fundraising New Year’s Eve celebration; I want classy, vibey, upscale fundraising party music that makes you feel like you’re at an upscale lounge, not overpowering, with jazz vibes, for a 60-second theme song.” I’ll click create — it’ll take a minute.

This is how everything ties together, going back to the beginning of the call and Mark’s question about how you take this stuff and make an actual video out of it. This is my workflow for turning this into an actual promo video. I’d normally spend more time locking in each slide, but for now: I like the second music option better, so I’m downloading that track. I’ve opened up Descript (genaiu.com/descript to get started). I find it’s easier to build the video over the music by uploading the audio first and then creating the images on top of it.

Once the MP3 audio downloads, I add it to the Descript project as the music track — now I have a roughly three-minute music track I can overlay other content on. There are a couple other things I want to add: the intro clip and the VO3 video generated earlier, and the actual slides from Gamma. While the slides aren’t animated yet individually, we could take each image and run it back through Gemini to generate video from the image. I’ll export the Gamma deck as PNGs, which exports all the slides together as a zip file.

This is like a crash course in getting a Descript video made: you upload your assets to the project, add the audio file at the bottom, add a new layer on top so it filters in, and then add your visuals in sequence. You can adjust the sound level of any video clip if you want it less overpowering. I recorded a quick “Happy New Year everyone” clip and will probably shorten it to under 5 seconds.

Once the zip file of exported slide images downloads and extracts, I drag all ten images into the project in sequence — this is the order Gamma generated them in, following the storyboard, which is why it helps to have that order locked in from the outline stage rather than generating a ton of unordered images and having to figure out sequencing after the fact. I tighten the timing on each slide — around 3 seconds each — so the final video comes out just over a minute. You could easily condense further.

Look at what we’ve created in this hour: using Gemini to create outlines, running it through Gamma using Nano Banana Pro to generate the images, generating the music vibe in Suno, and VO3 for the initial video clips. A question came in: did the presenter find that Gamma using Nano Banana Pro generated better pictures than Nano Banana directly inside a Gemini chat? The answer: not entirely sure if Gemini’s Nano Banana is running the same underlying model as Nano Banana Pro (there’s a flash version and a pro version), but generating images directly in the Gemini chat wasn’t following the request very well and was inconsistent, whereas going through Gamma with Nano Banana Pro generated images cohesively, following the style guide, in one shot — suggesting there’s likely something happening behind the scenes on Gamma’s end.

The group then showcased using Descript to tie everything together, including the theme song created for the event. One participant mentioned just signing up for Descript to try it. The host demonstrated recording a voiceover directly inside Descript — selecting “record audio recording,” speaking the voiceover live, and having it insert into the project and transcribe automatically. One feature highlighted: Descript’s Studio Sound tool on the AI tools panel, which can clean up audio recorded on a phone to sound more like it was recorded on a proper microphone.

The host noted this was getting a bit too far into a dedicated Descript class and that they were out of time for the day, but summarized Descript as a solid tool for video editing, letting you edit like a Google document — pulling in videos from Gemini or VO, Gamma slides, and music, and bringing it all together in one piece.

In the remaining Q&A: a participant asked about an affiliate link for Descript — the host pointed to genaiu.com/descript, noting a Black Friday deal that had actually already expired and would need updating. On plan pricing, the host is on Descript’s Creator plan (cited around $35/month) and Gamma’s Pro plan (cited around $20+/month, which includes Studio access). On Google’s VO3 usage limits, the host mentioned hitting a video generation cap during the session and using GenSpark — described as a broader agentic tool similar to Manus that also provides access to VO3 and other video models — as a workaround for extra V3 access. A participant also asked about an affiliate link for Whisper Flow (the voice dictation tool used throughout); the host said they didn’t have one yet but would look into becoming an affiliate.

Closing out, participants shared that the session was helpful and that they hadn’t previously realized all these capabilities (image generation, video generation, editing) were available directly inside Gemini rather than requiring separate tools. The host noted Gems (Gemini’s custom-GPT equivalent) hadn’t been covered and could be a topic for a future workshop, and pointed attendees to the Gen AI University Skool community for future live workshops.