Lisa Avvocato.

“You're going to learn a lot faster, and you're going to get more confident in your skills. The faster you can learn, the more relevant your skills are going to be.”

The before.

Lisa works as a fractional CMO and go-to-market advisor for early-stage SaaS companies, mostly pre-seed to Series B. Many of her clients are AI-native, so the topic was already in her daily orbit. The question was whether she could build with the tools, not just talk about them.

She tried for about a month on her own. The simple use cases, "like, you know, having an email manager or an autoresponder for a form", were not going to move the needle for the brands she advises. She wanted ad generators, blog systems, real workflows. So she went straight at the complex stuff and got stuck.

A week and a half on webhooks.

One of the worst rabbit holes was a week and a half spent trying to wire an Airtable form to n8n via webhooks, because that is what the tutorials told her to do. She isn't a coder. The whole detour disappeared the moment someone pointed her at the Airtable trigger node. "I was like, well, yeah, that would probably be a lot easier now, wouldn't it?"

That is the cost of trying to learn rapidly-evolving software in public. The tutorial was already six months old. The tooling had moved. She was paying the gap with her time.

What we built.

Lisa's first shipped build was an ad generator for a B2C client. The workflow side is solid. She is holding back from rolling it into live brand work because the image models still produce one or two tells that her trained eye picks up. That is a quality bar choice, not a build problem. She has worked too long in AI not to spot a hallucination.

Alongside the ad generator she built a blog content generator and an SEO strategy workflow. The SEO one ran into scraping reliability issues, which is now a recurring conversation. The build philosophy across all three is the same: sweat the prompts, let the technical layer get easier as the tooling commoditises.

Ad, blog and SEO workflows.

"You can have the technical setup. If your prompt sucks, you're going to get the AI slop that everybody talks about. I think that's a lot of where the value is going to be is somebody that really knows how to write and structure prompts."

Next on the list is a PR pitch tool that generates personalised outreach to journalists and podcast hosts for the brands she represents. She has just been given access to the n8n AI builder and is rebuilding the same tool inside it as a comparison.

The shift.

Lisa's bet is on prompt engineering as the differentiator, not the technical wiring. As the n8n AI builder and similar tools push the build layer towards drag-and-describe, the operators who can write the prompt structure that makes the agent actually useful are the ones who win.

You're going to learn a lot faster, and you're going to get more confident in your skills. The faster you can learn, the more relevant your skills are going to be.

The shift for her isn't time saved on a single workflow. It is the speed at which she can now learn a new pattern, ship something testable, and stay current in a field where everything she learned six months ago risks going stale. "I'd be concerned if I was trying to learn on my own, that everything I'm learning is becoming outdated."

Why this works.

Lisa got the most value out of the on-demand modules and the looms-back-and-forth feedback loop. As a parent of two with no patience for stuck-on-webhooks weekends, the format had to fit her schedule. "You'd sit and help troubleshoot with us, even if it was just online sending looms back and forth. It wasn't necessarily you had to be on a call." That flexibility, paired with someone slightly further ahead on the curve, was the thing that broke her solo plateau.

I'd be concerned if I was trying to learn on my own, that everything I'm learning is becoming outdated.