Search Everywhere: Why Finder's GM Killed His SEO Team

Very few people have run SEO as a line item and signed off on the budget that funds it. Zak Ali has done both. He started in search in 2015, co-founding the news blog Rant and growing it to roughly 100,000 readers on rapid-response articles and long-form think pieces. In 2018 he joined Finder, the editorially independent financial comparison site, as a publisher, then head of growth, and now General Manager of Finder US, where he owns organic, paid, email, CRM, and sales.

That vantage point has made him unusually blunt about what SEO has become. On a recent Unscripted SEO conversation, he argued that the job most SEOs are defending no longer exists, and the one that replaced it is bigger.

SEO didn't shrink. It got absorbed.

The framing that's caught on with practitioners, including AJ Kohn's "surface optimization", is that SEO now sits inside a much larger discipline. Large language models didn't carve out a slice next to search; they wrapped a shell over the top of it that also reaches social, video, and every other place a person can find an answer.

LLM really is what took it from search engine optimization to search everywhere optimization. Now it's really about being on every discoverable surface.

In practice, you can't coast off one surface anymore. Finder now runs what Zak calls "content waterfall" techniques: a written piece becomes a YouTube video, which gets cut into shorts and reels. The written word is the first domino now, not the finished product.

He dissolved the SEO team on purpose

Last May, he shut down Finder US's dedicated SEO team and folded it into a generalist growth org. It wasn't easy; those people had spent years building the discipline. But the logic was simple.

There are two types of SEOs that exist today. There are the technical SEOs, and there are SEOs who haven't realized yet that their job is digital PR and branding.

Editors, in his model, stop being editors of written content and become spokespeople, "niche messengers." A publisher is a marketer whose job is to be wherever the customer is, not to work Google for traffic in isolation.

Retention is the KPI nobody puts on the dashboard

If the surface count is exploding, the economics of chasing each new one get worse. Zak's warning here reframes what an SEO team is for.

The traffic you have today is as cheap as it's going to be. It's only going to get more expensive. Retention should be a core KPI of every SEO.

He's watched Google Discover and News traffic behave like a firehose that shuts off without warning, "very ephemeral." The teams winning there capture the visit into something durable instead of chasing raw impressions. He points to publishers gating even free content behind an email so an ephemeral moment becomes a relationship. For any SEO, that makes the job audience acquisition and cultivation, not just rankings, whether you sell brokerage accounts or, in his example, concrete walls in Florida.

AI is putting the customer first again

Zak pushes back on the idea that "slop" arrived with LLMs. In his read, homogeneous content is what SEO already incentivized: everyone reverse-engineering the P1 result until every article looked the same. The 2023–24 Helpful Content updates were the correction, and the canary was searchers appending "reddit" to escape the sameness.

What AI is doing is actually allowing us to put the customer first again. It's forcing companies to deliver more — and the customer actually wins.

For a YMYL business like Finder, that customer-first bar is non-negotiable anyway: ratings backed by a documented methodology, editorial independence stated in your face, and legally-mandated ordering (a lender comparison can't be sorted by who pays the most). Those "baked-in trust signals" are exactly the kind of thing AI systems are learning to reward.

The terminal-native research stack

Where Zak gets most animated is workflow. Finder US has moved the team into Claude Code, calling their SEO tooling and Google Workspace over MCP so keyword research and large-scale analysis happen without leaving the terminal. He's blunt that the tooling has flattened:

You don't need to pay thousands of dollars for Profound or Otterly. You can build your own LLM tracker with a free DataForSEO tier and AI. Use these tools for direction — then think like your customer.

His weekly team trainings are now about building a "second brain" in Obsidian: dumping your entire SEO process, decision log, and context into markdown so any model can run your playbook. From there he's built a "publisher" agent that does keyword research, customer profiling, the brief, competitor and deep research, internal linking, and the push into the CMS end-to-end.

This is the shift aeotrace was built for. The terminal-native workflow works because the data stopped being locked inside thousand-dollar suites, the same flattening Zak is describing. aeotrace is usage-based, and it ships an MCP server so an agent can run keyword research, inspect SERPs, and read your Search Console data from inside Claude Code, with no dashboard and no context-switching.

Keyword research didn't die, it got directional

None of this means the fundamentals evaporated. Asked which pre-LLM strategies survived, Zak's answer was measured:

Keyword research still exists as a directionally helpful tool. People don't really search in keywords anymore — they ask their question directly. But knowing your ICP, technical SEO, schema, speed — all of that still matters a lot.

The inputs haven't changed; you're attacking them from more angles. That's the premise behind our Keyword Research Strategy Library: seed demand from real conversations, then validate and expand it in the tool, rather than starting from a volume report.