In 2020, I realized the face of search was about to change radically. Not incrementally, radically. The signals were everywhere for anyone who had watched search evolve for decades: language models were getting genuinely good at understanding intent, and the gap between “ten blue links” and “just give me the answer” was closing fast.
By 2022, everyone was throwing AI at every project they could think of. It was the gold-rush phase, demos everywhere, substance rare. Most of what shipped in that period was a chatbot bolted onto a product that did not need one.
In 2023, when I was asked whether we should be investing in AI development, I said yes. Not because of the hype, despite it. The fundamentals underneath the noise were real: machines were becoming a primary *consumer* of web content, not just an index of it. That changes what a website has to be. I was right.
By 2025, Gemini, Claude, and to an extent GPT had matured into useful business tools. The gimmick phase was over; the fad had burned off. What remained was infrastructure, reliable enough to build on, accountable enough to put in front of clients. It was time to deliver real results, not demos.
And that is why custom-built Evergreen Optimization systems are now essential. Search is no longer one channel with one rulebook. Your customers find businesses through classic rankings, AI-generated answers, voice assistants, and local discovery, four consumers of the same content, each reading it differently. Off-the-shelf tools were built for the old world: one language (usually English), one channel (usually Google), one size for everyone.
One size no longer fits all.
This series is the proof, told through one real project, development began in December 2025, a custom, Hebrew-first SEO engine, codename Project Spark, built for a multi-location client, spanning traditional SEO, answer-engine optimization, generative and local discovery, AI workflows with human oversight, and the unglamorous engineering (byte-fidelity, testing, security) that makes it all trustworthy. Everything in these twenty-one posts happened; only the names are withheld, because my client’s competitive advantage is part of what I was hired to protect.
If you take one idea from the whole series, take this one: the businesses that win the next decade of search will not be the ones with the newest tools. They will be the ones whose content is engineered, deliberately, defensively, and durably, for every machine that now reads it.
Start with Part 1, Why I Built a Custom SEO Plugin Instead of Using Yoast.
