A week and a half ago I had a product and no real infrastructure to sell it.

No cold outreach running. No email capture. No nurture sequence. No paid ads. Just an offer page and some organic content.

By end of the week I had two separate funnels live. One organic, one paid. I want to walk through both — because the way I built them is, I think, more interesting than what I built.

Who I am and why that matters

I’m not a developer. I run a podcast, I’m building a productised AI consulting business, and I homeschool three kids in Byron Bay. I’ve worked in and around tech for years — angel investing, early-stage startups — but I don’t write code and I don’t have time to watch 45-minute tutorials every time I hit a wall.

That used to mean I either outsourced the work or it didn’t get done.

That’s changed.

Funnel one: organic

The organic funnel runs on two tracks simultaneously — the podcast and cold email via Apollo.

The podcast brings in listeners who already know me. The cold email (targeting Australian accounting firms via Apollo) brings in people who’ve never heard of me. Both tracks needed somewhere to land people that wasn’t just “book and pay $999” — too big an ask for cold traffic.

So I built a lead magnet. The AI Assessment I ran on my own business — a full bespoke report — is now available for free at jakewoodhouse.io/self-assessment. You enter your email, you get the report. The logic: let cold prospects read the actual deliverable before deciding whether to pay for one.

Building it meant getting into the WordPress backend for the first time, building a page in Elementor I’d never used, hosting a PDF on my own domain, and wiring a ConvertKit form to deliver the report automatically on signup.

Every time I got stuck — and I got stuck constantly — I took a screenshot, sent it to Claude, and got a specific answer to my specific problem. Not a tutorial. Not a forum thread from 2019. An answer to what I was actually looking at, right now.

Once someone signs up they enter a five-email nurture sequence I built in ConvertKit. Claude helped with the copy arc (welcome → proof → objection handling → guarantee → booking invitation) and the automation logic — how to tag subscribers, trigger the sequence, and graduate them to my weekly newsletter once they finish.

The moment that nearly broke me: getting emails to actually land in inboxes rather than spam folders. That meant editing DNS records at the domain level — settings that affect everything, including my business email. Touch the wrong one and you break it all.

Claude walked me through exactly which three records to add. And then told me specifically which existing record not to touch, and why. I made the changes. Nothing broke. Emails started delivering.

That “don’t touch this one” instruction is the kind of thing you’d spend hours finding on your own, if you found it at all.

Funnel two: paid

The organic funnel warms people up slowly. The paid funnel is a different bet — put money behind a specific message, target a specific audience, and see if cold strangers take action.

Target: Australian accountants, one to ten staff, partner or director level. Platform: LinkedIn, because that’s where they are. Budget: $50 a day.

I built three ad variants in Canva — each testing a different hook — with Claude helping on copy and strategy. Then set up the campaign in LinkedIn Campaign Manager: audience filters, budget, schedule, UTM-tagged URLs so the traffic is isolated in Google Analytics.

The paid funnel doesn’t ask for $999 upfront. That’s a big ask for someone who’s never heard of me. Instead it sends people to a dedicated landing page — jakewoodhouse.io/ai-assessment/accountants — with a single CTA: book a free 15-minute discovery call. Lower commitment, puts the conversation first, lets me do what I’m actually good at.

Which meant I also needed to build that landing page. Back into WordPress, back into Elementor, designing a page I hadn’t built before — with Claude helping me figure out why the two-column layout wasn’t rendering correctly and why there were accidental green highlights behind the button text.

Is the page perfect? No. But it’s live, it’s tracking, and it’s running.

What this workflow actually looks like

It’s not one big session where you hand the keys to an AI and walk away. It’s forty small moments of being stuck, describing what you’re looking at, and getting unstuck.

“How do I make these two columns sit side by side?” Screenshot. Done. “This URL looks wrong — is it going to 404?” Paste it in. Found the problem. Fixed it. “The automation tags aren’t firing.” Walk me through your logic. There’s the issue.

The other piece that makes it work is structure. I keep a master document in my Claude project files — business plan, product roadmap, decisions, all in one place. Every sub-thread I open for a specific task inherits that context. The prompts get sharper because the AI already knows what I’m building and why.

The point

One in three Australian small businesses say they don’t know where to start with AI (Deloitte, 2025). I’d argue the blocker for most isn’t willingness — it’s not having a clear picture of what practical daily use actually looks like.

This is what it looks like. Two funnels, ten days, no technical background. You describe the problem. You share the screenshot. You follow the next step.

If you want someone to map out specifically where AI could do this for your business — where the time is leaking and what the ROI looks like — that’s exactly what my AI Assessment does. Twenty-minute conversation, bespoke report within 48 hours, full refund if the identified monthly ROI doesn’t exceed the cost. Click here to learn more .