You can now generate custom PPC tools in plain English. With GPT-5 enabling complete program generation, the competitive edge belongs to those who master AI-assisted automation.
Frederick Vallaeys is building tools in minutes, not days or months, with AI. Vallaeys spent 10 years at Google building tools like Google Ads Editor, then another 10 building tools at Optmyzr, where he’s CEO.
He’s watched automation evolve firsthand, and vibe coding is the next leap. At SMX Next 2025, he shared his journey with vibe coding.
The traditional script problem
If you work in PPC, automation has always been top of mind. In the early days, you relied on Google Ads scripts. Scripts are great because there’s always more work than fits in a day.
But here’s the problem: when Vallaeys asks who actually writes their own scripts, only three to five out of 100 raise their hands. Most people copy and paste scripts because they don’t know how to code.
This works, but it’s limiting. You’re stuck with what someone else built instead of implementing your own secret sauce.
GPT changes the game
A couple of years ago, GPT made it easy to write scripts without knowing how to code.
The best part? Large language models are multimodal. You can take a whiteboard flowchart of your campaign decision tree, give the image to AI, and it’ll write the full Google Ads script.
Vallaeys suggests rethinking meetings. Instead of seeing client meetings as more work, treat them as prompt-engineering sessions.
It’s easy to get frustrated when clients add more to your plate. But with a mindset shift, the meeting becomes the prompt that tells AI what to execute.
What is vibe coding?
Instead of writing lines of code, you describe what you want the software to do, and the AI handles the technical implementation. That’s vibe coding.
Imagine your team needs software that does X, Y, and Z. Write down what it needs to do, give it to a coding tool, and it builds the software. As Vallaeys says, it’s mind-blowing.
Scripts are old news. Vibe coding is the new frontier.
A live example: Building a persona scorer
Vallaeys showed how fast this works. He went to Lovable and said, “Build me a persona scorer for an ad that shows how well it resonates with five different audiences.”
In less than 20 seconds, the AI responded with its design vision, features, and approach. It explained exactly what it would build, so he could immediately say, “Actually, make it 10 audiences instead of five.”
You work with it like a human developer — without touching code. You just describe what you want changed.
The framework: What should you automate?
Traditionally, you automated two types of work: quick, frequent tasks (like reviewing search terms) and long, infrequent tasks (like monthly reporting with analysis).
Vallaeys advises you not to limit automation to what you already do. Think about what you wish you could do more often but haven’t because it’s too time-consuming. That’s prime automation territory.
The old way vs. The new way
The old process was painful. Launching something took at least a month.
You’d spend days writing specs. Engineers would spend days building. You’d find bugs, coordinate meetings, and repeat.
The other problem? Traditional code was deterministic — pure if/then logic. Great for reliability, but terrible for nuanced decisions like, “Is this a competitor term?” It’s nearly impossible to program every variation of competitor keywords.
The promise of on-demand software
Sam Altman announced GPT-5, leading with “on-demand software generation.” The industry is moving beyond software-as-a-service to true on-demand software.
The new way? Write a one-paragraph spec (five minutes), give it to AI (15-minute build), then review and iterate (three minutes per change). In under an hour, you have working automation.
This new code is flexible, not just deterministic. LLMs can answer nuanced questions like, “Is this a competitor term?” with high probability. It’s the best of both worlds.
The expanding scope of automation
With vibe coding, anything you can explain to a human, a machine can build. Landing pages that follow brand guidelines? Done. Custom audience tools? Done.
Here’s the radical shift: you can now automate tasks that take just 90 minutes by hand. Build throwaway software for one-time tasks. Even if it breaks next month, it saved you time today.
What can you build with vibe coding?
You can build landing pages, microsites, interactive web apps, Chrome extensions, browser extensions, and WordPress plugins — all through simple prompts.
Start with Claude or ChatGPT — tools you likely already subscribe to. They’re great for data analysis, calculators, and quick visualizations.
For more complex apps that need databases or login systems, use Lovable, V0.dev, Replit, or Bolt. They handle the complexity, so you don’t have to.
If you’re more technical, try Codex, Bolt.new, or Cursor. But for most people, the simpler tools handle almost everything.
Vallaeys asked someone on his team who had never coded to build a seasonality analysis tool. She fed PPC Town Hall podcast videos into Claude.
The process was simple: gather resources, write a prompt, give it to AI, and test it in the browser. No installation required.
The team iterated on the fly, asking for different plots and forecasting methods. In minutes, they had advanced enhancements. The AI knew where to add help text and simplify the interface because it’s trained on millions of web apps.
Vallaeys wanted multiple custom GPTs to review his blog posts in sequence, each giving feedback from its persona. Then a consolidator GPT would summarize the most common feedback into three to five bullet points.
He vibe-coded this in V0.dev by describing what he wanted. It generated a clean tool with text input, the ability to add custom GPTs, and everything worked.
Case Study 3: Chrome extension for demos
For customer demos, Vallaeys needed to blur sensitive numbers. He wanted options:
- Fully redact or just blur?
- Include currencies or only numbers?
- Handle different separators?
He built a Chrome extension with all those options using simple prompts. Problem solved.
Prompting tips for success
Always include the use case. Say “seasonality tool” instead of vague terms like “time series analysis.” The AI makes better assumptions and may suggest approaches you hadn’t considered.
Ask questions: “How did you approach this?” or “Where do you store data?” It helps you learn.
Use chat mode to explore alternatives without changing the code. Ask for three approaches, pick one, go deeper, then say, “Execute that.”
The PPC audience analyzer
The audience analyzer Vallaeys’ team built is available to try. You can grab the code, add your logo, turn insights into action items — whatever you need. Just tell it what to change, and it updates.
Final thoughts: Stay competitive
Vallaeys makes one point clear: you’re not competing against AI. You’re competing against people who use it better than you do.
Try vibe coding today. Go to one of these tools and give it a single prompt. See what happens. The first time Vallaeys tried it, his mind was blown.
Now that you’ve learned something new, use it to get better at AI. That’s how you stay ahead.
Search Engine Land is owned by Semrush. We remain committed to providing high-quality coverage of marketing topics. Unless otherwise noted, this page’s content was written by either an employee or a paid contractor of Semrush Inc.
