Results · my own portfolio
This is my largest business: a home-improvement publishing business that also sells house plans and design guides. Readers come for ideas about kitchens, exteriors, layouts and finishes. The site earns from the audience it builds, from its email list, and from the plans and guides it sells.
AI took over the content, the social pages, the Facebook ad buying and the email list. Here’s what happened, chart by chart.
Owned and operated by Jon Dykstra · not a client
Jon's own portfolio. Figures are from businesses Jon Dykstra owns and operates on AI. They are not client results. Data to Oct 11 2026 unless noted. The site is described by niche, not named.
A publishing business lives or dies on volume and judgment at the same time. You need a lot of good pages, a lot of reach and a lot of tested ads, and you need someone deciding every day which of them deserve more money.
For years the limit wasn’t ideas. It was hours. A person can write a few pages a week, check a handful of campaigns and post when there’s time. In a business this size, that ceiling was costing traffic every week.
In late June the business was doing 256,000 sessions a week. Once the AI systems ran together, traffic climbed every month. The week of October 5 it did 609,000 sessions and 759,000 pageviews.
AI writes, designs and schedules posts across 22 Facebook pages and makes more of whatever people stop scrolling for. Organic views went from about 3,800 a week in June to 75,500 a week in October, with no ad spend behind them.
In February, AI was managing US$460 a week in Facebook ad spend. AI now writes the ads, launches them, sets bid caps and judges every campaign every day. Spend only grows where the tracking proves it should: US$10.8K a week in October, US$60.6K in 2026 and 1.45 million ad clicks.
AI researches, writes, illustrates and checks every page, then launches its ad campaigns the minute it goes live. Output went from 35 pages a week in June to 107 a week in October: 2,772 pages in 2026 and 708 ad campaigns launched automatically.
The email list sits at about 49,000 subscribers with a 46% open rate. AI-tested sign-up offers added 1,141 sign-ups in the first 13 days they ran (Sep 30–Oct 12 2026), and home-equity lender leads come in on top of that.
Every morning the system ties yesterday’s spend to yesterday’s results for every live campaign, then applies the rules: pause, hold or scale. The rules themselves get tested. An early version judged new campaigns too soon and switched off some that would have become winners. Testing the rule against real results showed the problem, so the system now gives campaigns more data before it judges them, and campaigns that would have been switched off stay alive and profitable.
Content works the same way. The system reads which topics, titles and formats earned the most attention and uses that to choose and shape the next batch. Every lesson is written down, so nothing has to be learned twice.
Jon's own portfolio. Figures are from businesses Jon Dykstra owns and operates on AI. They are not client results. Data to Oct 11 2026 unless noted. The site is described by niche, not named.
Book a free 30-minute call. You'll leave knowing exactly what AI should take over first in your business and what it should be worth to you.
Only a small number of businesses are taken on at a time.