The rebuild
About 3–4 hours, as reported in the session, guided by a plain-language brief. The 3–4 hours describe this specific client project.
Case study · NY City Blinds · Agency client
The task: replace a heavy WordPress site with fast pages ready for paid traffic. The AI worker did the rebuild with approved GitHub and Cloudflare access while Adam Argaman’s team directed and reviewed each step.
This video couldn’t load. Open the recording or read the transcript below.
It means: okay, I give you the current website. I want you to rebuild it. And I did that in a few hours, three, four hours.
And I did it in a few hours, and this website you built a few years back in the day.
One of the problems with the Google Ads account is that we're getting in some campaigns very low ad quality scores, so some of our ads are not showing up, and for the ads that are showing up, we're paying more because the page is not optimized.
It can be the site speed optimization, because the current website is sitting on WordPress, so you have a lot of components and plugins.
But you don't really have a dedicated landing page that you can send the traffic to. And if you send the traffic, everything is loading very slowly, and they need to move between pages and the navigation and all of that.
We want, as part of using an AI worker and Outloop infrastructure, to do things with the help of the AI agent.
I was able to create and duplicate the whole website, with photos, with images, with CRO, conversion rate optimization, for the website, with an implementation for CRO, like PostHog. So it's reading the traffic, and you can do the optimization in the right place.
So we have here, on the official website of Outloop, under Guide: AI Managed Landing Page.
Basically, we're using GitHub to host all the files, with permission to deploy when you push to main. Everything is detailed here. And Cloudflare. Cloudflare is dealing with the DNS and with R2.
You go to Outloop, connect the API key, as we know, and that's it. You have all the capabilities.
After I gave the agent access to GitHub and Cloudflare, basically I gave him all the abilities to be our developer, to do everything on the page.
This is the full website. The thing is that everything is extremely optimized for Google Ads.
And just for comparison, you will see the existing website. This is the existing website.
WordPress. Very heavy. The resolution, you can see the resolution is bad. People cannot read this, like, very wide resolution.
Look at the recommendations.
Look how we're showing the stores. We show them with images, all the locations, Google Maps, SVG animation. And the most important: the Calendly is everywhere. Every landing page, the Calendly is there.
So I started to create a document, just in natural language.
And then, because I want the client to be happy and approve this, move fast.
So think like a conversion rate optimization specialist, and think about the right positioning for this section.
But try not to remove sections that we've created, just reorganize them and reduce unnecessary content, because we believe it's better to visualize with SVG icons and images. We have a lot of images, and if you don't find the right image, you can generate them with OpenAI Image, or the most recent one, and create something authentic. If you use people, use identical people. If you use the logo, use the identical logo, but you can change the positioning and create the best image per section.
Try to keep the content similar to the official website, so the client can approve the copywriting faster.
And I asked him: "Okay, are we ready to support a Google Ads campaign?" Like, we have all the tagging, we have CallRail. CallRail is like a dynamic phone number. Each user from a different traffic source will see a different number. And then we can identify them by UTM and by source.
So he tells me: almost, but not yet. I can see green on CallRail, phone number, Tag Manager, remarketing, GA4, Meta Pixel, Bing, Pinterest, Taboola, Outbrain. Almost everything, except the lead conversion. Why? Because we're working with Calendly.
For Calendly you need a thank-you page. It has no thank-you page.
Fix the phone number in the footer to match the CallRail number, and report to me when we are ready to support traffic from paid ads, especially from Google Ads.
So now Yoni, the AI worker, just sent me a message, after I told him to fix Google Tag Manager and all the recommendations over here. So let's see what he just sent back. I like to see everything green. So basically, we're ready for paid traffic. He verified.
Okay, he went to Google Tag Manager. I remind you, we asked him to create a fresh new thank-you page and add the pixel for all the platforms in Google Tag Manager, and he just did it in 11 minutes.
This is the thank-you page. Nothing fancy. We can obviously tell him to do this and that. But everything is ready.
Wow, Yoni, great work. Thank you so much. For the thank-you page, I think it will be a great strategy to warm up the leads, so they can know more about us, about our story, about the people behind, about our showroom.
So figure it out. I trust you. Do the work. Send me back the thank-you page for my review, and just report to me again that we are ready to support paid traffic.
We pay a lot. We need to optimize and improve the Google Ads strategy. What you need to do: you need to create all of these pages." And the client freaked out. He didn't know where to start, because he's very busy with a million other things. And this is the old way. And then we came with this strategy for the client, so the AI worker will have access to do everything.
And then I asked him: "Okay, if we implement all of this strategy, give me your thoughts, your estimation: how much better can we be? What is the percentage that we can potentially improve the Google Ads campaigns?" So he came to me with a very interesting statistic. Yoni, the AI worker, is very conservative, but he gave me some nice statistics here.
So he's basically saying between 20 and 90% more conversions on the same budget.
So if we go to the middle, around 50%, it's crazy.
So what we want to do here: we also work on the CRM. We connect the CRM deals from Zoho CRM to Google Ads. So now we'll be able to see the actual sales, and not only the leads, which is really great for Google Ads, because you can teach the Google Ads algorithm to track those qualified sales and deals.
The problem is that it's new. And now we're going to ask the AI worker to activate a Routine, starting from the beginning of November, to track every week when is the right time to do the shift of the primary conversion, to basically train Google Ads to follow those actual deals and value, instead of just leads.
He will never forget. He will remind us, and he will decide what is the best timing.
Also, Yoni, we have one of our best, not one of, our best campaigns: the location campaign. So go directly into Google Ads, understand the location campaigns. We need to have a dedicated landing page to send traffic to those location campaigns. Figure it out, use your design skills, think like a CRO expert, and build a dedicated landing page per location, or per campaign. You will figure out what the best practice is here.
What I want you to focus on first: our active campaigns on Google Ads, and make sure we have an absolutely CRO-optimized landing page for each one of our campaigns, and get back to us with the results.
And yeah, that's it. That's the power of an integrated AI worker. They've been part of our internal team, and you can communicate with them. You can be your developer, your CRO expert, your content writer, your Google Ads and Meta Ads expert, your design studio expert. Everything, in a session of one hour.
About 3–4 hours, as reported in the session, guided by a plain-language brief. The 3–4 hours describe this specific client project.
Asked whether the site was ready for paid traffic, the worker found one gap, the lead conversion, and added a thank-you page and the pixels in Google Tag Manager in 11 minutes.
Through Outloop, the worker had approved access to the GitHub repository and Cloudflare. The team set the goal and reviewed every step.
Times are as reported in the session, not independently audited. No conversion or cost results are claimed; the “20–90% more conversions” figure in the video is an AI-generated scenario estimate, not a measured result. Every client’s stack is different, so integrations are scoped before setup.
A 30-minute walkthrough with the Argaman X team. We’ll look at the task, the tools involved and what a useful result would be.
Book a Demo