From Freelancers to Agents-as-a-Service

For the last decade, outsourced creative services have been defined by cost advantage.
Why pay agency rates when you could hire globally, get work delivered fast, and keep budgets lean?

But the cracks are everywhere.
Freelancers compete on price, not originality. Agencies pad hours with executional tasks anyone could do. Most “outsourced creativity” isn’t creativity at all. It’s production at scale.

That model is now collapsing.
Fiverr just laid off 30% of its workforce and is pivoting to AI services. Execution work is no longer a safe business to be in. What was once offshored is now being automated. Freelancers see it first. Traditional agencies will see it next.

Catching up means seeing the shift clearly.
The market for hands-on production is dying. The market for orchestration, quality control, and brand judgment is just beginning. Instead of buying tasks, companies will buy systems: agentic creative workflows run by people who know how to steer them.

So what do you do?

If you’re on the client side:
Audit your outsourced spend. Strip away every task that’s pure execution. Reinvest in strategic direction, taste, and oversight. Test AI-first outsourcing models now—before your competitors normalize them.

If you’re an agency:
Stop billing for execution. That revenue stream is gone. Rebuild your value around orchestration, creative direction, and judgment. Position yourself not as “hands on keyboards” but as the system designer who ensures AI output is on-brand, original, and effective.

The future of outsourced creativity isn’t cheap labor.
It’s AI as the worker, humans as the director.

— Peter & Torsten

Before we jump in, a word from this week’s sponsor:

Still guessing which campaigns actually work? Discover how leading brands cut waste and scale smarter with incrementality. This free guide breaks down how to measure true impact, optimize spend, and move beyond outdated attribution. Download now and start making smarter, data-backed marketing decisions.

Luma AI has released Ray3, its new flagship generative video model (you can try it here). The key upgrade isn't just better visual quality or 4K HDR output. It's inference. Unlike previous models that directly translated prompts into pixels, Ray3 introduces a reasoning layer.

It analyzes a creative request, breaks it down into logical steps like shot planning and lighting, generates the video, and then evaluates its own output against the original intent, refining as needed.

This new architecture has a "Draft Mode," which lets creative teams iterate on concepts up to five times faster and more cheaply than before. Once a direction is approved, the shot can be "mastered" into a production-quality file.

The model also introduces a new level of control with visual notes, allowing users to draw directly on a frame to direct camera movement or character placement, significantly reducing the need for complex prompt engineering.

The shift from generative to "inferential" video radically changes production workflows. Instead of locking in a concept and hoping the AI gets it right, teams can now treat AI as a junior creative partner, rapidly exploring ideas in a low-cost draft mode.

The ability to create and test dozens of video ad variants, with different moods, settings, or character actions, becomes economically viable.

YouTube has rolled out a suite of powerful new AI-driven features for its YouTube Studio platform, designed to streamline content creation, provide deeper audience insights, and enhance brand safety.

A new "Inspiration Tab" acts as an AI-powered content strategist, analyzing a channel's performance and audience engagement to recommend new video ideas. To provide deeper insights, a new chatbot assistant called "Ask Studio" allows creators to ask natural language questions about their account, such as consolidating viewer feedback from comments or analyzing engagement trends.

On the brand safety front, a new "likeness detection" tool will automatically identify and flag unauthorized use of a creator's image, helping protect key figures like a company's CEO from deepfakes or misuse.

YouTube also announced new generative AI tools specifically for Shorts, including a custom version of Google's Veo 3 Fast model and an "Edit with AI" feature to simplify short-form video creation.

These updates embed AI directly into the core workflow of the world's largest video platform. The Inspiration Tab is especially useful: content strategy moves from manual guesswork to data-driven recommendations.

Furthermore, the "likeness detection" tool provides a critical, automated first line of defense against reputational damage from AI-generated misinformation.

Meta has unveiled its new Ray-Ban Display smart glasses, and the standout feature for marketers is "Live AI." This new capability allows an onboard AI assistant to see what the user sees through the glasses' camera and hold a real-time conversation about it.

While similar functionality is available today through smartphone apps like Google Lens or ChatGPT's vision features, the hands-free, low-friction interface of smart glasses is designed to make this behavior habitual.

This technology allows a user in a retail store, for example, to look at a product and ask the AI contextual questions like, "Is this a good choice for a small apartment?" or "What are the key differences between this and the model next to it?" The AI can access real-time information to provide an immediate answer, effectively bypassing traditional search engines at the critical moment of decision.

Live AI creates a new, high-intent marketing channel that sits directly between a customer and a product in the physical world. This is the closest we've come to providing a perfect, on-demand salesperson to every customer, everywhere.

Just as companies are scrambling to ensure their products appear in LLM text responses, they must now focus on making their real-world products and services "legible" to visual AI assistants.

OpenAI is about to roll out a native checkout feature directly within ChatGPT. A hidden "Orders" section was discovered in the latest desktop app update, allowing users to track purchases initiated and completed inside the chat interface. The initial rollout will likely be limited to the U.S. market.

Customers asking an AI for purchase advice represent one of the highest-intent audiences available, second only to someone walking into a physical store. The rollout of in-chat purchasing creates a powerful new conversion channel that bypasses traditional search and e-commerce funnels. The time to ensure your brand is visible and transactable in AI responses is now.

Online freelance marketplace Fiverr announced it is laying off 30% of its workforce as part of a strategic pivot toward AI-driven services. In a statement, CEO Micha Kaufman emphasized that the company is adapting to a reality where AI can perform many of the tasks previously handled by human freelancers.

The company's strategy now involves integrating AI tools and potentially offering AI-generated services directly to customers, fundamentally changing its business model from a human-centric marketplace to a tech-enabled creative platform.

The announcement was met with significant backlash from the creative community, many of whom have already reported a steep decline in job offers and income due to businesses opting for cheaper, faster AI alternatives.

Fiverr's pivot is a canary in the coal mine for the creative services industry. The value proposition of external agencies is moving away from pure execution (which can be automated) and toward strategy, creative direction, and quality assurance.

This will force a re-evaluation of agency relationships and in-house team structures, prioritizing those who can orchestrate AI tools.

Small Headlines. Big Shifts.

» Google's Gemini Gems Are Now Shareable

Google now allows users to share their custom AI assistants, known as Gems. This update makes it easier for teams to create and distribute specialized AI agents for specific tasks, like analyzing sales data or drafting on-brand social media copy.

» An Economy for AI Agents

Google announced the Agent Payments Protocol (AP2), an open standard designed to allow AI agents to pay each other for services using stablecoins. This foundational step could enable complex, multi-agent workflows where specialized AIs contract with each other to complete tasks.

» Notion 3.0 Puts AI Everywhere

Notion has launched version 3.0, deeply embedding AI into the core of its workspace platform for all users. New features include a natural language global search and "AI Autofill" for databases that automatically summarize and structure information.

» Zoom's AI Drives Toward a 3-Day Workweek

Zoom CEO Eric Yuan predicts that AI digital twins will automate so much work they could enable a 3-day workweek. Meanwhile, the company rolled out AI Companion 3.0, with new features to summarize meetings, analyze customer conversations, and automate workflows within the Zoom platform.

» Video Avatars Reach Uncanny Valley

Both VEED and Kling AI showcased significant advances in video avatar technology. VEED's Fabric 1.0 and Kling's latest demos feature hyper-realistic digital humans with natural expressions and movements, showing that high-quality, scalable AI-presenter videos are becoming a viable option.

AI is dismantling the old economics of creative work. The outsourcing model (cheap labor at scale) isn’t coming back.

The leaders who thrive will be the ones who act now: cutting executional bloat, rewiring workflows, and investing in orchestration and taste.

This is where the next competitive edge is built. And no CMO should have to figure it out in isolation.

That’s why we built the AI Ready CMO community: a space where marketing leaders learn, experiment, and adapt together.

Apply for your membership: join.aireadycmo.com

See you next week,
Peter & Torsten

PS: Know someone who would benefit from this newsletter? Refer them and get rewarded: