Luma Launches AI Agents That Handle End-to-End Creative Work
Luma's new AI agents, powered by Unified Intelligence models, can execute complete creative workflows across text, images, video, and audio. Already used by Publicis, Adidas, and Mazda to slash production costs.
Luma Launches AI Agents That Handle End-to-End Creative Work
Luma Launches AI Agents That Handle End-to-End Creative Work
AI video-generation startup Luma has launched Luma Agents, a new class of AI collaborators that can execute complete creative workflows from initial brief to final delivery across text, images, video, and audio.
The agents are powered by Luma's new Unified Intelligence family of models, starting with Uni-1 — a single multimodal reasoning system that can "think in language and imagine and render in pixels," according to CEO Amit Jain.
What Makes Luma Agents Different
Most creative AI tools require constant back-and-forth prompting. You generate an image, tweak the prompt, regenerate, tweak again. Luma Agents take a different approach.
Instead of prompting for each iteration, the system generates large sets of variations and lets users steer the direction through conversation. The agents maintain persistent context across assets, collaborators, and creative iterations — and can evaluate and refine their own outputs through iterative self-critique.
"You need that ability to evaluate your work, fix it, and do that loop until the solution is good and accurate," Jain told TechCrunch. This self-correction capability is what has made coding agents so effective, and Luma is now bringing it to creative work.
Unified Intelligence: One Model, Multiple Modalities
The Uni-1 model is a decoder-only autoregressive transformer operating over a shared token space that interleaves language and visual reasoning. Unlike systems that bolt together separate models for text, images, and video, Uni-1 was trained on audio, video, image, language, and spatial reasoning from the ground up.
Luma calls this "intelligence in pixels" — the ability to maintain a unified mental representation of creative work, similar to how an architect visualizes a building as they draw.
The agents can also coordinate with other AI models, including:
- •Luma's Ray 3.14
- •Google's Veo 3 and Nano Banana Pro
- •ByteDance's Seedream
- •ElevenLabs' voice models
Real-World Results: 99% Cost Reduction
In one demonstration, Luma Agents turned a brand's $15 million, year-long advertising campaign into multiple localized ads for different countries — in 40 hours for under $20,000.
The output passed the brand's internal quality controls and accuracy checks.
Another demo showed how a 200-word brief and a single product photo (a tube of lipstick) led the system to generate various ideas for locations, models, and color schemes for an ad campaign.
Who's Already Using It
Luma has rolled out the platform to major customers:
- •Publicis Groupe — Global advertising agency
- •Serviceplan — European agency network
- •Adidas — Sportswear brand
- •Mazda — Automotive manufacturer
- •Humain — Saudi AI company
"Our customers aren't buying the tool; they're redoing how business is done," Jain said.
Availability
Luma Agents are now available via API, though the company plans to roll out access gradually to ensure reliable service and avoid workflow disruptions for users.
What This Means for Creative Teams
The creative industry has been waiting for AI to move beyond single-asset generation to full workflow automation. Luma Agents represent a significant step in that direction.
For ad agencies, marketing teams, and design studios, the implications are substantial:
- •Faster iteration: Generate and evaluate dozens of creative directions in minutes
- •Lower costs: Dramatic reduction in production expenses for multi-market campaigns
- •Consistent context: Agents remember brand guidelines and project requirements across iterations
- •Self-improving outputs: The critique loop means better results without constant human intervention
Bottom Line
Luma's entry into the AI agents space targets one of the most AI-ready industries: advertising and marketing. With major agencies and brands already on board and documented cost reductions exceeding 99% in some cases, this is a launch worth watching.
For creative professionals, the question isn't whether AI will handle more of the production pipeline — it's how quickly they'll adapt to orchestrate these systems rather than compete with them.
Sources:
Share this article
About NeuralStackly team
Expert researcher and writer at NeuralStackly, dedicated to finding the best AI tools to boost productivity and business growth.
View all postsRelated Articles
Continue reading with these related posts
Microsoft Copilot Tasks Launches in Research Preview: AI That Can Actually Do Your To-Do List
Microsoft Copilot Tasks Launches in Research Preview: AI That Can Actually Do Your To-Do List
Microsoft has launched Copilot Tasks in research preview, positioning it as a consumer-facing AI agent that can browse the web, manage recurring tasks, coordinate across apps, a...

Cursor's Fast Regex Search: How AI Agents Can Search Massive Codebases Without Waiting
Cursor built a local sparse n-gram index to replace ripgrep for agent search, eliminating 15+ second grep latency in large monorepos by pre-filtering candidates before full rege...
OpenAI Codex Computer Use Review: Background Agents That Control Your Entire Mac
OpenAI Codex Computer Use Review: Background Agents That Control Your Entire Mac
OpenAI Codex now controls your Mac in the background with computer use, scheduled tasks, 90 plugins, and an in-app browser. Full review of the April 2026 update.
Microsoft 365 E7 Launch: Microsoft Bundles Copilot and Agent 365 Into a New $99 Frontier Suite
Microsoft 365 E7 Launch: Microsoft Bundles Copilot and Agent 365 Into a New $99 Frontier Suite
Microsoft has launched Microsoft 365 E7, a new enterprise bundle that combines Microsoft 365 Copilot, Agent 365, and its security stack into a single $99 per-user plan aimed at ...
Claude Sonnet 4.6: Anthropic's Mid-Tier Model Now Matches Flagship Opus at One-Fifth the Cost
Claude Sonnet 4.6: Anthropic's Mid-Tier Model Now Matches Flagship Opus at One-Fifth the Cost
Anthropic's Claude Sonnet 4.6 delivers near-Opus performance across coding, computer use, and agentic tasks while costing 80% less. The new default model features a 1M token con...