Your Keyword Tool Can't See This

Wave 288; Mt. AI Crumbling; AIO Impacts on CTR; GEO is Reputation Problem; 3 Web Eras; Bottom-up Funnel; Amazon Ads Boycott; and Much More!

Author

Yoyao
April 25, 2026 • Est. Learning Time: 11 minutes

In partnership with

FIRST …

I've been wanting to fold AI search responses into Floyi’s Topical Research for a while. The thinking was always there. The execution kept getting away from me.

This latest update finally pulls it off with AI Search Gaps. This one I love.

AI Search Gaps generates personalized prompts from your buyer persona, runs them across ChatGPT, Gemini and Google AI Mode, and fuses what comes back into your topical research.

That direction matters. SEO, GEO and AEO are all moving the same way: more personalization. Google already knows everything about you. ChatGPT, Claude and others are getting there fast.

Search is becoming personal to whoever's typing. Your topical research has to do the same.

That's where Floyi has been heading from day one. Every map is shaped by your brand and your buyer personas because anything less produces a list of terms anyone in your space could pull. Most SEO keyword tools still serve up exactly that.

The few that track AI search prompts take a keyword, capitalize the first letter, slap a question mark on the end and call it a prompt. Something like:

  • "What are the best topical map tools?"

Floyi runs prompts like:

  • "I run SEO at an agency with 20 retainer clients across different verticals. Which tools can build topical maps that match each client's taxonomy without senior-led deep dives every time?"

Same topic. Completely different answer.

AD

The summit for the marketers of the future

Manual marketing had a good run.

But the teams winning right now aren't briefing, approving, and repeating. They're directing AI agents that execute the whole strategy for them.

The Agentic Marketing Summit (May 4–8) is a free, five-day event that shows you exactly how it works in practice. Not theory. Not a PDF checklist. Step-by-step insight to help you become an expert in AI marketing agents.

Hosted by 3x Inc 5000 founder Manick Bhan alongside the sharpest minds in the marketing world today.

The era of doing it yourself is over!

SEO + GEO

In this Search Off the Record episode, Gary Illyes and Martin Splitt walk through how Google used HTTP Archive data, custom metrics and BigQuery to analyze robots.txt directives at scale instead of relying on assumptions or a few anecdotal examples. Rather than arbitrarily document one or two unsupported robots.txt directives, the team wanted data on which unsupported rules and typos actually appear most often across the web. The conversation gets technical, covering how Chrome UX Report data feeds HTTP Archive, how WebPageTest and custom JavaScript metrics can extract robots.txt patterns, and how BigQuery can then be used to surface the distribution of directives, errors and malformed files. Read the transcript.

Glenn Gabe documents a case where a site scaled heavily with AI-generated local news content, received a manual action for scaled content abuse, then saw the expected collapse in Google’s index and AI surfaces like AI Overviews and AI Mode. What makes the post more interesting is the follow-through into ChatGPT, where Glenn found that many citations to the penalized directory also disappeared once the content was removed from Google’s index. He provides another practical example of how risky, low-quality scaling tactics can bleed across search and AI visibility instead of staying confined to one platform.

My Take: This is one of the clearest cautionary examples for anyone still thinking they can separate “AI search tactics” from classic search quality signals. If your content strategy is weak enough to trigger a manual action, there is a good chance the fallout will not stop at Google rankings. Long-term AI visibility is much more likely to come from stronger content systems than from scaled loopholes that work until they don’t.

Danny Goodwin reports on Liz Reid’s latest comments about how AI is changing Search behavior, with users asking longer, more natural-language queries and using Google more often because the system can respond to fuller intent instead of just keyword shorthand. Reid’s framing is that AI Overviews filter out low-value “bounce” clicks while still sending people to the web when they want to go deeper, and that users do not want AI or the web, but both together. She also says AI slop is not a new problem so much as a scaled version of an old one, and that Google’s job is still to surface the best information while monetization adapts to more detailed, potentially more commercial queries.

My Take: This is one of the clearest statements yet about where Google wants the narrative to land. AI is not replacing the web, and users are not leaving search. They are just expressing intent more naturally and expecting Google to do more of the translation work. That means the old model of matching simplified keywords is getting weaker, while the ability to answer fuller, messier needs (more personalized) is getting more important.

Reza Moaiandin explores whether original, high-quality content actually performs better in traditional search and AI citations, then lands in a much messier but more realistic place than the usual “just create better content” advice. His team’s qualitative research found some relationship between originality and performance, especially on queries that require interpretation or judgment, but not one strong enough to say that quality alone consistently wins. The most useful part of the piece is the case study showing how a fairly ordinary page targeting an emerging keyword like “API design” outperformed stronger competitors simply by getting there first and owning the topic before the market caught up.

My Take: I like this because it pushes back on the lazy idea that better content automatically wins if you just keep polishing it. Sometimes the real edge is spotting the opportunity earlier, being more original in what you choose to cover, or moving fast enough to own a topic before everyone else realizes it matters. Great content helps, but it is not some magic ranking bean.

Tracy McDonald, Hannah Cooley and Marketa Williams analyze 53 brands, 5.47 million tracked queries and 2.43 billion organic impressions to understand how AI Overviews are affecting organic and paid CTR in 2026. AI Overviews still suppress clicks, but some of the decline appears to be stabilizing, with several segments outperforming the team’s downtrend projections.

  • Organic CTR on AIO-present queries rebounded from 1.3% in December 2025 to 2.4% in February 2026, an 85% increase in two months.

  • Organic CTR when no AIO was shown rose from 2.8% in January 2025 to 3.8% in February 2026.

  • Informational queries showed AIOs 36% of the time, compared with 8% for commercial and 5% for transactional queries.

  • Comparison queries triggered AIOs 95.4% of the time, and question-format queries triggered them 85.9% of the time.

  • Being cited in an AIO delivered 120% more organic clicks per impression than being present on an AIO SERP without a citation.

  • On informational queries, average organic CTR was 2.07% when the brand was cited in the AIO, 0.94% when it was not cited, and 3.35% when no AIO appeared.

  • Transactional queries with AIO and brand citation doubled organic CTR from 0.7% in January 2025 to 1.7% in December 2025.

  • Paid CTR on AIO-present queries stayed relatively stable, while organic CTR remained far more volatile.

My Take: This is exactly why broad “AI Overviews kill clicks” takes are too blunt to be useful. You need to investigate your cases much more deeply: what kind of query it is, whether your brand is cited, and whether Google decides the query even deserves an overview in the first place.

Barry Schwartz reports that Google has now published best practices for the “read more” deep links that can appear within search result snippets, giving site owners a clearer path to making those extra snippet links eligible. The guidance is simple: keep the target content immediately visible instead of hiding it behind tabs or expandable sections, avoid JavaScript that forces the scroll position on page load, and do not strip the hash fragment from the URL if you use History API calls or hash modifications.

Gaetano DiNardi says that most GEO advice overstates the value of technical hacks and formatting tricks because LLM visibility is driven more by brand reputation, category alignment and external validation than by isolated on-page tweaks. His core case is that many popular tactics, from FAQ stuffing to markdown-heavy formatting and listicle chasing, are mostly table stakes or even distractions when the real issue is whether AI systems consistently understand and trust what category your brand belongs in. He also makes the useful distinction between citations and recommendations, showing how brands can earn page-level citations without becoming the brand AI systems actually recommend for a category.

Dan Taylor explains that content quality still matters in AI search, but no longer guarantees visibility because the real bottleneck has shifted from authorship and ranking toward retrieval, citation and distribution across a wider ecosystem. His point is that content now has to function not just as something worth reading, but as something easy for AI systems to extract, reuse and validate against other signals.

My Take: Great content still matters, obviously, but it is not enough when so much goes into accessibility, retrieval, and authority for AI Search (and all that goes into SEO for traditional search engines too).

SEO + GEO Ripples

AI

OpenAI is launching workspace agents in ChatGPT, Codex-powered shared agents that can run long workflows in the cloud across team tools, memory and approvals rather than acting like one-off personal assistants. Users can turn repeated internal processes into reusable agents that gather context from systems like Slack and connected apps, follow team rules, ask for approval on sensitive steps and keep running even when no one is watching. OpenAI is also framing them as an evolution of GPTs, with support for shared use across organizations and research preview availability in Business, Enterprise, Edu and Teachers plans before credit-based pricing begins in May.

Chris Talbott recaps Google Cloud Next ’26 as a broad enterprise AI push built around agents, infrastructure and security rather than a single product launch. The biggest themes are the Gemini Enterprise Agent Platform, the Gemini Enterprise app, new eighth-generation TPUs, the Agentic Data Cloud, Wiz-powered security layers and deeper Workspace Intelligence integrations.

My Take: Not much from a consumer view, but if useful for enterprises to know what’s going on.

Josep Garcia introduces a free AI toolkit guide for small businesses that bundles 26 tools across common workflows like writing, image creation, video, automation, research and lightweight building tasks. Small teams and solo operators are already doing a bit of everything, so this helps reduce time spent experimenting blindly.

Ivy Levine rounds up several Gemini use cases that go beyond generic prompting, especially when paired with your phone camera and Gemini Live. She walks through ideas many people probably have not considered, like showing Gemini a cluttered drawer for storage suggestions, panning across your fridge to turn leftovers into meals, pointing the camera at a broken appliance for repair help, checking a plant’s environment visually, and using room photos with Nano Banana to test design changes before moving furniture or repainting.

My Take: I haven’t even thought about some of the Gemini uses here. Maybe you already knew, but I never thought about using the camera on food for meal plans.

AI Ripples

  • OpenAI releases GPT-5.5, its smartest general model yet, built to take on messier multi-step work across coding, research, data analysis, documents, spreadsheets, software operation and early scientific workflows without giving up speed.

  • Google released Gemini Embedding 2, their first natively multimodal embedding model, mapping text, images, video, audio and documents into one shared embedding space for retrieval, classification and search tasks.

  • DeepSeek has previewed new V4 AI models that it says nearly close the gap with frontier systems, pairing a 1 million-token context window with lower pricing and stronger reasoning benchmarks while still trailing top models somewhat on knowledge tests.

  • Google published 1,302 real-world gen AI use cases from organizations across industries. It’s an idea library for teams trying to see where AI already shows up in operations, marketing, support, analytics and product workflows. If you need business-use inspiration, browse the deployments here instead of starting from a blank page.

  • Google open-sourced Stitch’s DESIGN.md format, a draft spec carrying design rules across tools so AI systems understand why colors, components and patterns exist instead of just copying their appearance. You can get it through their GitHub.

  • OpenAI launched an ads manager and cut ChatGPT’s ad pilot minimum to $50,000, lowering the barrier for advertisers while making its ad business look more like a real self-serve platform instead of a high-touch experiment. OpenAI is building the infrastructure it needs if it actually wants ChatGPT ads to scale into a mainstream media-buying channel.

  • Google is reportedly investing $10 billion in Anthropic, with another $30 billion potentially tied to milestone targets. They’re deepening a strategy where cloud and chip providers back external AI labs that may also become major infrastructure customers.

MARKETING

Tim Frank says marketers now need to operate across three parallel web environments at once: the human web where people search and compare manually, the LLM web where AI helps users choose, and the emerging agentic web where AI agents may complete the transaction for them. Microsoft uses that framing to introduce a set of product updates around Clarity AI Visibility, UCP-ready merchant feeds, Copilot Checkout, AI Max for Search, Offer Highlights and AI-assisted audience generation, all designed to help brands become more visible and actionable inside AI-driven decision paths. They say that agent traffic is rising much faster than human traffic, which means structured data, machine-readable commerce inputs and AI-aware measurement are becoming necessary infrastructure instead of optional experiments.

Annie Palmer reports that hundreds of large Amazon sellers organized a 24-hour boycott of Amazon’s advertising platform to protest a cluster of policy changes that they say are squeezing already thin margins. The changes include delayed seller disbursements, changes to how ad payments are collected, and a temporary 3.5% fuel surcharge tied to higher energy costs, all of which merchants say threaten cash flow more than simple profitability. The large organized seller community that collectively represents billions in revenue.

Jason Barnard outlines a marketing approach built around how AI tools are reshaping discovery, qualification and conversion, with the funnel starting closer to product value and workflow usefulness than top-of-funnel brand storytelling. He emphasizes that AI-assisted buying journeys can compress research and reduce the importance of some classic awareness-stage tactics, which puts more pressure on product-led proof, clearer use cases and lower-friction entry points. Rather than treating AI as a traffic gimmick, it frames the opportunity as redesigning acquisition around how people now move from need to action with AI helping them along the way.

My Take: This is a good read because a lot of marketing teams still act like the funnel starts where it did five years ago. AI compresses research and comparison, so the strongest acquisition systems may be the ones that get users into value faster instead of spending too much energy on awareness-stage content.

CONTENT

Kimberly Zhang explains that audio is one of the clearest ways to fight content decay, especially for strong content assets that spike briefly after publishing and then fade while audiences move on to passive, ears-open consumption. She lays out 11 repurposing ideas, from narrated blog posts and audio snacks to private podcast newsletters, Q&A responses and audio roundups of industry news, all aimed at turning written assets into something people can consume while commuting, exercising or multitasking.

My Take: Number 11 is one that speaks to me directly because it’s about roundups like this newsletter. So…do you want an audio version of this roundup?

SOCIAL MEDIA

Alex Juel walks through how to move YouTube analytics into Data Studio so teams can build cleaner, more shareable reporting without relying on direct access to YouTube Studio. The guide covers setup options like whether to use Google’s template or start from scratch, how to handle permissions when you are not the YouTube account owner, and which template errors need to be fixed before the report is actually usable.

Social Ripples

  • X has launched the standalone XChat app on iOS, splitting messaging out into a separate product with private chats, disappearing messages, file sharing, and audio and video calls while trying to give Communities users a new destination after shutting that feature down. My feeling is just like this:

TOOLS AND RESOURCES

Email Detective is a Chrome extension from EmailTooltester that identifies which email service provider sent a message inside Gmail and surfaces related authentication details directly on the subject line. The extension says it can detect more than 110 ESPs, including platforms like Mailchimp, Brevo, HubSpot and Klaviyo, while also showing SPF, DKIM and DMARC status, possible shared versus dedicated IP usage, and a new bulk scanner for checking up to 50 emails at once. That makes it a lightweight competitive research and deliverability inspection tool for anyone who spends time reverse-engineering email stacks.

My Take: This is a very cool tool and gives you all sorts of infrastructure data. If you do any email marketing competitor research, it’s a tool you’ll want.

WAYS WE CAN WORK TOGETHER

Floyi - Build Topical Authority that wins in Google and AI Search. Don’t just plan your content strategy - make it unstoppable.

TopicalMap.com Service - Let us do the heavy lifting. We handle the research, structure, and strategy. You get a custom topical map designed to boost authority and dominate your niche and industry.

Topical Maps Unlocked 2.0 - Unlock the blueprint to ranking success. Master the art of structuring content that search engines (and your audience) love - and watch your rankings soar.

AD

Speak the email. Send the email.

Talk through your reply and get polished, professional text ready to paste. Wispr Flow strips filler, fixes grammar, and formats everything. 89% sent with zero edits. Works everywhere.

LIKE DIGITAL SURFER?

Find me and others in the Digital Surfer Discord community.

I’d also love to know what you think and if you have any ideas for the newsletter. Reply or email me at [email protected].

I’d also appreciate it if you shared it with fellow digital surfers.

You currently have 0 referrals, only 3 away from receiving LinkedIn Shout-out.

Have a great week taking your SEO and digital marketing to another level!

And don’t forget to drag the Digital Surfer emails to your Primary Inbox 🌊 

Reply

or to participate.