Sara & Nik’s Podcast:
Julian Galluzzo

About this conversation

Sara Taher and Nik Vujic, CEO of Get Stuff Digital, welcome Julian Galluzzo in today’s Sara & Nik SEO Podcast episode to discuss SEO, website development, and how AI has crept its way into the field.

The topic of the discussion is AI’s impact in web development. In short, AI can code functional websites, but are they good to the standard required in 2026? This is where Julian’s expertise comes in, who, as someone with experience in agentic development, can provide valuable insight in how AI agents work and how and where they fall short.

Julian’s Path from Webflow Loyalist to Expert Agentic Builder

Julian’s tech journey started when he was just 13 and he’s learned everything online. Over time, he realized that he didn’t like to do things the standard way. He would always look for shortcuts – ways to speed up the process without sacrificing quality along the way.

He discovered this trait of his as he moved from running a web agency to working in the Webflow community, to then occupying a role at Memberstack, a company built around Webflow. This inclination of his towards speed and efficiency is what eventually brought him to AI, more specifically GPT-3.5.

First, he was using it to write small scripts that would otherwise require a developer. Those scripts turned into automations, which, eventually, grew into tools like Cursor. It was at that point that Julian realized that he hadn’t opened Webflow in months – working with AI was faster, more intuitive, and the quality was consistent.

So, he did the only reasonable thing he could do: he migrated Memberstack’s site off Webflow. The problem was that the workload was now scattered, since it was no longer housed within Webflow’s workshop. This was the main discontent within the Webflow community, among the people who’ve watched Julian’s video announcing the transition.

Fortunately, he came up with a solution: Ship Studio, his own proprietary platform designed specifically to pull the agentic workflow in one place. An open source platform that gives web professionals precisely what they need: a faster and more organized workflow.

What AI Can Do and Why It Still Needs Humans

As Julian puts it, AI agents (LLMs in particular) are nothing more than prediction engines. Given a task, the AI can predict the next logical step to create a working code – an area where, he believes, AI has already beaten humans.

But this enhanced capability is also one of AI’s biggest weaknesses, because, as Julian explains:

“I'm yet to see an agent autonomously set an actual good and inspired design direction.”

The AI can take the steps to reach a given goal, but cannot set the goal itself. It needs a human to do that for it. This brings up an even more interesting point: the idea that humans are still a vital part of the process. What changed is their position within the system – the human operator no longer occupies the bottom position designed to coding, the AI has taken over that. Instead, the human is now in charge of oversight and end-product polishing.

“You need to know what a good outcome looks like in order to manage a person or an agent” explains Julian, forecasting an era which Nik calls “the era of operators”. In other words, AI has made content and audits cheap and easily manageable, but humans are still needed to assess and correct the output.

The Problem of Accountability and the Future of SEO

The problem that transpired the strongest during the discussion was: is AI development ready for high-traffic websites in 2026? Julian believes so, with one condition:

“It's as ready as the people who are managing it are.”

This raises the problem of accountability and the idea that, if a client’s site performs poorly or flat-out goes down, nobody really blames the AI model, they blame the people at the helm, the experts in charge of handling the AI workflow. Nik mirrors the exact sentiment by explaining that, no matter the level of automation, he always has the final say with regards to the end project’s quality.

Unfortunately, the direction that the development process is heading is one born out of a misunderstanding of what AI is and what it does. Too many experts outsource too much of the work to AI with not enough supervision.

From Nik’s point of view, this can only result in one thing: websites built with AI from grounds-up that never get indexed. And the trend is, unfortunately, already underway.