You have a new vertical to launch, zero extra headcount, and a tight deadline. Your first instinct is to throw one massive prompt into an AI tool and hope for the best—but that always fails. In this session, Daniel Zikry from Natural Intelligence reveals how his team built a fully automated AI workflow by treating AI like a team of specialist agents, taking a brand-new vertical from zero to top-spending ROI from day one.
The “Magic Prompt” Myth
When tasked with launching a completely new vertical with no additional headcount and a hard deadline, performance creative teams are often forced to look for a shortcut. For many, that shortcut looks like a “mega-prompt”—asking an AI tool to brainstorm, write, design, and resize an entire campaign in one click.
As Daniel Zikry explains, this approach fails completely. AI is not a magic button. When you treat it like one, you get generic, unusable outputs. The secret to scaling a creative team that produces over a thousand assets a month isn’t about writing one perfect prompt; it’s about treating AI like a team you manage.
Building Your AI Creative Team
To build the “One-Click Ad” workflow, Daniel’s team at Natural Intelligence stopped asking a single AI model to do everything. Instead, they broke the work down into specialist AI agents: one for creative strategy, one for copy, one for design, and one for video.
Here is how they built a workflow that actually converts:
1. Feed Your AI the Right Diet
An AI agent is only as good as the context it receives. Instead of asking AI to “write a good ad,” you must feed it high-quality, real-world data. Daniel’s workflow ingested:
- Transcriptions of competitor videos and top-performing competitor ads.
- Real, unfiltered quotes and pain points from Reddit and Twitter.
- The company’s own historical top-performing creatives.
- Specific pack shots and rigid brand kits.
2. The R&R Method: Role, Responsibility, and Objective
If you want an AI pipeline to stay focused, you have to manage it like a human employee. Daniel introduced the R&R Method for prompting. For every AI agent in your pipeline, you must explicitly define:
- Role: Who is the AI? (e.g., “You are a senior direct-response copywriter.”)
- Responsibility: What is their exact task? (e.g., “Write three distinct hooks based on this Reddit research.”)
- Objective: What is the ultimate goal? (e.g., “To stop the scroll and drive clicks from CFOs looking to cut software costs.”)
3. Separate Planning from Execution
A major mistake teams make is asking AI to ideate and execute simultaneously. Daniel stresses the importance of separating the two. First, use your AI agents to build out the strategy and create the “Masters”—the core, hero concepts of your campaign. Only after the Masters are approved should you trigger the execution agents to build iterations, A/B testing variants, and multi-platform resizes.
The 5 Golden Rules of AI at Scale
While AI can help you produce thousands of assets a month, maintaining quality requires strict guardrails. Daniel closed his session with five golden rules for implementing an AI creative-at-scale workflow tomorrow morning—proving that with the right architecture, you can achieve massive scale without sacrificing creative integrity or ballooning your payroll.
🎥 Watch the Full Session
Want to see the exact AI agent architecture Daniel built, or learn the five golden rules for scaling your own creative pipeline?
Watch the full video to see how Natural Intelligence built the “One-Click Ad” workflow and turned a stressful deadline into an ROI-positive launch from day one.
This article is based on a session from Creatives and Conversions 2026 (Creatives & Conversions TLV III), brought to you by the Growth Marketing Pros community. Stay tuned for more cutting-edge strategies from top industry experts.
