Luma's new AI agents produce text, images, video, and audio from a single brief. Here's what multi-format AI content production means for SMBs.
Article text
A marketing manager at a 40-person commercial cleaning company told me something last month that stuck with me. She said, "I spend more time switching between tools than I spend actually creating content."
She wasn't exaggerating. For a single campaign promoting a new service, she was using Canva for graphics, ChatGPT for copy, a stock video service for B-roll, a separate tool for voiceovers, and then manually stitching everything together in a video editor. Five tools, three days of work, one social media campaign that ran for a week.
That workflow is about to look very different.
AI Content Production Just Went From Single-Tool to Full-Workflow
Earlier this month, a company called Luma launched AI Agents that handle end-to-end creative production across text, images, video, and audio. Not four separate tools. One agent that takes a creative brief and produces multi-format content autonomously.
Their system, powered by what they call "Unified Intelligence" models, coordinates multiple AI systems in a single workflow. You give it a brief. It plans the content, writes the copy, generates images, produces video, and creates audio. The agent can even pull in specialized models from Google, ByteDance, and ElevenLabs when a specific format needs a different engine.
Major ad agencies like Publicis Groupe are already using it. But the implications for small businesses are arguably bigger.
If you're a 20-person company spending real time and money on content production, we can help you figure out which parts of that workflow are ready for AI . Free 30-minute discovery call, no pitch.
The Difference Between AI Tools and AI Agents for Content
This distinction matters more than most people realize, and it's the key to understanding what just changed.
AI tools
do one thing per prompt. You ask ChatGPT for a headline. You ask Midjourney for an image. You ask ElevenLabs for a voiceover. Each request is separate. Each output lives in a different app. You're the project manager connecting them all together.
AI agents
do the whole workflow. You describe what you need (a product announcement for social media, with a short video, three static images, caption copy, and an audio snippet for a podcast ad), and the agent plans and executes across all those formats. You review the finished package instead of managing each piece.
For a marketing team of two or three people at a small business, this is the difference between producing two campaigns a month and producing eight. Not because the AI does everything perfectly, but because it eliminates the coordination tax: the time spent switching apps, exporting files, reformatting assets, and manually ensuring visual consistency across formats.
What This Actually Costs Today
Let me put some real numbers on this, because "AI will save you money on content" is meaningless without specifics.
A typical SMB content production setup might look like this:
Freelance graphic designer:
$500-$1,500/month for social media graphics and marketing materials
Copywriter (freelance or part-time):
$800-$2,000/month for blog posts, emails, social captions
Video production:
$1,000-$3,000 per video (or $200-$500/month for stock footage subscriptions)
Tool subscriptions:
$200-$400/month across Canva, stock photo services, scheduling tools
Internal time:
15-25 hours/week of a marketing person's time on production tasks
Total monthly spend on content production for a 30-50 person SMB: $2,500-$7,000, plus a significant chunk of someone's job.
AI content agents don't eliminate all of this. You still need a human making strategic decisions: what to say, to whom, and why. You still need someone reviewing output for brand consistency and accuracy. But the production work — the part where you're actually building the assets — compresses dramatically.
The companies we work with that have started using AI-assisted content workflows typically see the production portion of that spend drop by 40-60%. Not because they fired anyone, but because their marketing person went from spending 20 hours a week on production to spending 8, and redirected the rest to strategy, outreach, and customer conversations.
Three Content Workflows Worth Testing First
If you're running a small marketing team and want to experiment with AI content production, don't try to overhaul everything at once. Start with these three workflows, because they have the highest ratio of time saved to risk involved.
1. Social media repurposing
You wrote a blog post. Now you need five social media posts, three image variations, and a short video clip. This is the single highest-value use case for AI content agents because the thinking is already done. The strategy, message, and angle are set. You just need the content reformatted across platforms.
Start by giving the AI your blog post and asking it to produce platform-specific versions. Review and adjust for tone. This alone can save 3-5 hours per blog post published.
2. Internal communications and training materials
Most SMBs dramatically underinvest in internal content. New employee onboarding, process documentation, training videos. These always fall to the bottom of the priority list because they're time-consuming and don't generate revenue directly.
AI content agents change the math here. Recording a quick voice memo explaining a process and having an agent turn it into a formatted SOP document, a training video with visuals, and a quick-reference card takes a fraction of the time it would take to produce manually.
3. Seasonal campaign production
HVAC companies running winter heating promotions. Landscaping companies pushing spring cleanups. Accounting firms promoting tax prep services. These campaigns happen every year with minor variations. The creative brief barely changes.
Feed last year's campaign into an AI agent, tell it what's different this year, and let it produce the updated materials. Then spend your time on the strategic decisions — targeting, offers, messaging — instead of the production.
What AI Content Agents Can't Do (Yet)
I'd be doing you a disservice if I didn't flag the limitations.
Brand voice is hard to automate.
AI agents produce technically competent content, but matching the specific tone and personality of your brand requires iteration and human judgment. The first draft will be close. It won't be right.
Original thought leadership doesn't come from agents.
If your content strategy depends on genuinely novel insights (and it should), the thinking still needs to come from you. AI can produce content efficiently. It can't replace the perspective that comes from actually running a business in your industry.
Quality varies across formats.
AI-generated text is further along than AI-generated video. The images are good but sometimes need tweaking. Audio is getting there but can sound synthetic. Multi-format agents coordinate well, but each individual output is only as good as the underlying model for that format.
Consistency requires guardrails.
Without brand guidelines, style guides, and specific instructions, AI agents will produce something generic. The businesses that get the most value from these tools are the ones that invested in documenting their brand standards first.
What to Do This Week
If you're spending more than $3,000 a month on content production (including internal time), it's worth mapping your content workflow end-to-end.
Write down every step from "we decide to create something" to "it's published." Include every tool switch, every handoff between people, every export and re-import. Then circle the steps that are pure production — no strategic thinking required, just building assets.
Those circled steps are your automation candidates. Some of them are ready for AI content agents today. Some will be ready in six months. But if you don't know where the production time goes, you can't make an informed decision about what to automate.
We map content workflows as part of our broader process audits. If you want to see where your marketing team's time is actually going, and which pieces are worth automating now versus later, let's talk . We'll give you an honest assessment, even if the honest answer is "keep doing what you're doing for now."