--- title: 'I Built a Marketing Team with 1 AI Agent & No Code ๐Ÿค– ' source: 'https://youtube.com/watch?v=tBeCkyHPywg' video_id: 'tBeCkyHPywg' date: 2026-09-22 duration_sec: 489 channel: 'NexaAI_Studio' --- # I Built a Marketing Team with 1 AI Agent & No Code ๐Ÿค– > Source: [I Built a Marketing Team with 1 AI Agent & No Code ๐Ÿค– ](https://youtube.com/watch?v=tBeCkyHPywg) ## Summary This video demonstrates how a single AI agent, hosted on the NN platform and controlled via Telegram, can manage an entire marketing department. It routes text or voice commands to six specialized sub-agents that handle blogging, social posting, image creation/editing, video production, and more โ€” all for mere pennies per task. ### Key Points - **Central AI manager on Telegram** [00:13] โ€” A single AI agent lives inside Telegram, allowing users to chat via text or voice instead of using multiple software platforms. - **Master logic flow** [02:09] โ€” The central AI receives messages, transcodes voice notes, analyzes intent via a system prompt, and acts as a switchboard delegating tasks to six sub-agents. - **Blog and social post pipeline** [03:00] โ€” For a blog post request, the AI uses Tivoli for live research, writes the post, generates a matching graphic prompt, sends it to OpenAI, and returns a formatted post to the user. - **Image editing with memory** [03:54] โ€” The AI recalls past creations from a Google Sheets database and Google Drive file IDs, enabling context-aware edits via OpenAI's edit endpoint. - **Automated video production workflow** [04:53] โ€” From a simple prompt, the AI creates a storyboard, generates images via Flux, animates them in Runway, adds audio via ElevenLabs, and stitches everything in CreatorMate to deliver a final video. - **Video pipeline details** [05:30] โ€” Four storyboard scenes become image URLs, which are animated into 5-second clips, matched with custom audio prompts, and merged into a 20-second movie. - **Cost breakdown per task** [06:14] โ€” Image editing costs ~20ยข, Flux images ~1.5ยข, Runway animation ~25ยข, and CreatorMate ~1 credit. The monthly overhead for NN cloud hosting is just $27. ## Transcript Welcome to This Explainer. Okay, today we are unpacking an absolutely mind-blowing concept from the world of no-code automation. We are going to explore exactly how one AI manages marketing. And I don't just mean a few tasks. I mean how one single AI agent, hosted entirely on a platform called NN, can be engineered to run an entire marketing department. If you've ever felt totally overwhelmed by the sheer volume of content creation it takes to run a business today, well, you're going to absolutely love this breakdown. Just imagine for a second a world where you never have to log into five different software platforms just to get a single campaign out the door. Imagine if sending a simple, casual text message or, you know, even a quick voice note instantly commanded an AI to output deeply researched logs, stunningly realistic images, and fully animated videos. Sounds like magic, right? Well, it's not. It's just really, really smart routing. Okay, let's dive into this. Here's our roadmap. First, we'll meet your new AI marketing manager. Then, we'll peek at the brain of the operation, explore the blog, design, and video departments. And finally, we'll look at the bottom line on cost. All right, section one, meet your AI marketing manager. The true beauty of this system is its sheer simplicity on the user side. Your central AI manager lives entirely inside Telegram. There's no complex dashboard. You literally just chat with it using text or voice exactly like you would a human employee. Look, we all know the drill with a traditional marketing team. The problem is often bloat, right? It's expensive. There's a ton of slow back-and-forth communication. You get human bottlenecks. And you're paying for a dozen different software subscriptions. In stark contrast, this one AI setup operates on literal micro-costs, delivers almost instantly, and effectively manages six sub-agents all hosted under one roof in A-Nade-N. Let move to section 2 the brain of the operation So how does this central boss actually coordinate all these tasks without getting totally confused It all comes down to a master logic flow The basic logic of this central brain is actually pretty elegant Step one, it receives your telegram message. If it's a voice note, it transcodes it. So voice and text are standardized as the exact same type of data. Step two, the agent analyzes your intent based on a really clear system prompt. And step three, well, this is where the magic happens. This single AI acts as a master switchboard operator, instantly delegating your command to the exact right virtual employee for the job. And this brilliantly illustrates that setup. That single main agent manages six distinct virtual employees just sitting on standby. A blog writer, a LinkedIn poster, a graphic designer, an image editor, a video director, and even a database librarian to search past files. It is literally an entire agency just waiting on your word. Moving right along to section three, writing blogs and social posts. Let's step into this first specialized department. This isn't just about spinning up generic, boring text. This agent can research and write highly targeted copies. Picture this. You're grabbing your morning coffee and you just send this prompt. Create a blog post about the effect of sleep on productivity for working professionals. That's it. Within moments, the AI kicks into gear. Here is the exact pipeline that fires off. First, it uses a search tool and plugin called Tivoli to conduct live web research, gathering fresh data. Second, it writes the post, specifically tailoring the tone to your target audience, which, in this case, is working professionals. Third, it analyzes the blog it just wrote and creates a prompt for a matching graphic. And finally, it sends that prompt to OpenAI to generate an aligned image, returning a fully formatted post and graphic right back to your phone. Section 4 Designing and Editing Stunning Visuals. Now, creating an image from scratch is impressive, for sure, but the system's ability to recall and edit existing visuals That is where it truly shines Say you just generated an image of a cat for a flyer but it looks I don know a bit too cartoonish You simply type can you edit that image and make it more realistic So the crucial point is right here. The AI has memory. It understands the context of that image. It doesn't start from scratch. It knows exactly what you are referring to. So how on earth does it pull this off? Through a really clever background process. Basically, the AI acts as a librarian. It searches a Google Sheets database where it logs every single piece of content it creates. It finds your most recent generation, grabs a unique Google Drive file ID for that specific image, and passes that file to OpenAI's edit endpoint. A minute later, boom, your newly edited, highly realistic cat flyer pops right back into your Telegram chat. All right, section five, directing fully automated videos. This is the grand crescendo of the entire setup. It is a multi-step, incredibly complex workflow that gets triggered by something as simple as typing, create a video of a beaver building house. Because video requires structure, right? So the very first thing that original single AI brain does is act as a master visual story director. It takes your prompt about the beaver and rapidly splits it into a four-part cohesive storyboard, ensuring there's an actual logical flow from gathering sticks to placing them in the river. Next up, the system takes those four storyboard scenes and sends them simultaneously via an API to an image generation model called Flux. This spits out four distinct, highly detailed image URLs representing each beat of the Beaver story. But hey, still images aren't a video, so Step 3 automatically feeds those four image URLs directly into Runway, which is a super powerful AI video generation tool. Runway analyzes the images and the original prompts, animating them into four separate five-second video clips. And you know, a video without sound feels totally lifeless. So while those visuals are rendering a dedicated sound agent writes customized prompts for each of the four scenes and fires them off to 11 laps Incredibly it generates bespoke five immersive audio effects like the sound of rushing water or snapping twigs to perfectly match each specific video clip. Finally, we reach the finish line. All of these separate binary files, the four video clips, and the four audio tracks are routed into a tool called Create-A-Mate. Create-A-Mate automatically stitches them all together, perfectly timing the audio to the video and renders a final 20-second movie. The agent then downloads that completed file and delivers it straight to your Telegram chat. It is just a stunning display of orchestration from one single AI manager. Which brings us to section six, the bottom line on cost. Now, what's really interesting about this is that when you replace heavy software subscriptions and human specialists with API calls, the economics completely flip. Let's walk through the literal pennies it costs to run these massive creative tasks. Generating or editing a high-quality image with OpenAI, that costs about 20 cents. Generating images using Flux is a fraction of that, at roughly one and a half cents per image. Animating a five-second video clip in Runway, about 25 cents. And stitching it all together in CreatorMate costs about one account credit. When you run a full blog, image, or video command, you are spending pocket change per campaign. And beyond those tiny micro-costs for the API calls, and maybe a few cheap $5 base plans for tools like 11Labs, the primary overhead to run this entire digital agency is just $27. That is the monthly cost for the NAN cloud hosting, where the central brain lives. We've seen exactly how one AI manages marketing by intelligently routing text and voice commands to research blogs, edit realistic imagery, and direct fully animated short films, all for literal pennies. It fundamentally changes how we think about marketing leverage. So I'll leave you with this provocative thought. With your marketing department running entirely on autopilot from your phone, what will you have the time to build next? Thanks for joining me on this explainer, and keep exploring.