Duane Forrester has a striking comparison for AI-era SEO: it feels remarkably like the late 1990s, when he started. That comes from someone who has watched search for 25 years, a decade of it from inside Microsoft.
Forrester started in SEO before Google existed. There was almost no documentation, almost no tooling, and no direct channel to the engines. Learning meant running experiments, reading code, and trading discoveries with other practitioners in online communities and at conferences.
Today rhymes with that period because the rules are being written in real time again. The difference is an excess pointing the opposite way: back then there was nothing to read, now there is far too much.
This is where his advice gets practical. Stop trying to keep up with the leading edge. Unless you work inside Google, Microsoft, OpenAI, or Anthropic, you cannot see the true leading edge anyway. Aim for steady progress and keep the edge merely in sight.
Forrester's biggest surprise after joining Microsoft was how compartmentalized a search engine is. He asked someone on the Bing Webmaster Tools team for backlink data, and the employee asked back whether links actually affected rankings.
It was not a joke or a deflection. Links were not part of that person's job, and they had no access to that part of the system. It permanently changed how Forrester reads statements from engine representatives: working at Google or Microsoft does not mean understanding every part of search.
The same experience produced a second lesson. SEOs can obsess over tiny details with almost no practical impact. In his cooking analogy, stick with the chunky onions.
Forrester also pushes back on the idea that engines and SEOs are natural adversaries. When Google says to focus on users and their journeys, it is not purely corporate misdirection — engines are chasing the same customers marketers are.
That is why he argues for dropping the labels "technical SEO" and "content SEO." Technical excellence will not rescue content that misses user intent, and strong content struggles without the technical foundation needed to earn trust. Modern search demands both, which lines up with the reordering laid out in SEO Priorities for 2027.
The Google decision that still frustrates Forrester most is the removal of keyword referral data. He called it a slap in the face to the industry: useful data disappeared and businesses had to decide with less information.
There was an unintended effect. The data gap created demand for third-party SEO platforms, and in his view that shift is part of how SEO grew past its cottage-industry roots. He would still like the data back.
At Microsoft, a manager once told Forrester he was underperforming. He expected to lose his job. Then a lunch with industry veteran Stephan Spencer reframed the problem: it was not his ability, he was simply trying to hold too much work in his head.
He built a system that broke projects into subprojects, tasks, deadlines, and individual work items. Within months his performance improved sharply, and at his annual review he got a promotion and the largest bonus available instead of a dismissal.
The lesson is blunt. When the workload becomes overwhelming, the answer is not always to work harder — sometimes it is a better system. That applies squarely to marketing teams now absorbing AI tooling on top of their existing channels.
Asked what search looks like in 20 years, Forrester says agents. Instead of researching a coffee maker yourself, you tell an assistant what you need; it researches the options, applies what it knows about you, and returns a shortlist — eventually completing the purchase.
That does not mean search goes away. It becomes an infrastructure layer beneath those experiences, supplying the information agents need to understand and act on a request.
Content still matters. Trust matters more. Brands have to demonstrate trustworthiness to people and to machines — which is the supply side of the consumer shift described in They Closed the Search Box.
Three moves translate directly.
First, stop making "keeping up" the goal. A team that tries to review every AI search update burns out and never touches the big blocks. Pick two or three large items per quarter instead.
Second, treat engine statements as clues, not sources. Check whether the speaker owns that area and how far their visibility actually extends.
Third, remove the job labels from your org. When technical and content sit in separate seats, one of user intent or technical foundation is always left uncovered — and in AI search, that gap shows up as a citation you never earn.
The technology will change dramatically. The mission probably will not: understand what someone wants at the moment they want it, and be the best answer.
Forrester's advice is to stop chasing the leading edge. Outside Google, Microsoft, OpenAI, or Anthropic you cannot see it anyway. Pick two or three large items per quarter and keep the edge merely in view.
Check whether the speaker owns that area. Forrester's example is a Bing Webmaster Tools employee who did not know whether links affected rankings. Engines are compartmentalized, and employment is not omniscience.
Technical work cannot rescue content that misses user intent, and good content cannot earn trust without a technical foundation. When ownership is split, one of the two is always uncovered.
Forrester expects search to become an infrastructure layer rather than disappear, supplying the information agents act on. Content is still required, and proving trustworthiness to both people and machines becomes more important, not less.
To apply what you just read to your own site, start with a free audit of where things are now.
A strategist replies within 24 hours on business days.
The Google Search favicon bug is a week old and Google has confirmed it is on their end. Here is how to tell whether it is your site, and why this week's CTR drop should not enter your reporting as a result.
2027 SEO strategy shifts from ranking to being understood, trusted, and cited. Here are the priorities: AI answer engines, E-E-A-T, brand mentions, zero-click, and structured data.
If Google vanished tomorrow, how would your brand get discovered? Starting from that question puts distribution before owned content — and produces the external signals search and AI systems actually read.