Square Peg SEO, AI Agent Boards & In-Page Push Newsjacking (Brian Kato & Dan Kurtz)

Brand-saturation SEO, a Claude "board of directors" agent build, and unblockable in-page push newsjacking from the Day 2 Kato/Kurtz session.

This Day 2 session ran as two back-to-back workshops bracketed by a short hallway-style exchange. The opener is an unidentified premium-PR operator (name not given in the source) arguing with the hosts that a budget "PBN news network" of cheap press releases cannot match premium distribution on authority or brand-perception grounds. Brian Kato (Square Peg SEO) then presents a brand-saturation framework that abandons rank tracking in favor of owning the entire SERP and AI-answer narrative, and live-builds a "Board of Directors" of Claude AI agents that produces strategy decks and keyword/geo-grid spreadsheets. Dan Kurtz (In-Page Push / newsjacking) closes by teaching how to hijack other people's news sites with unblockable HTML push notifications and newsjack Google Trends into AI-generated forced-conversion landing pages.

Main takeaways

1. Rank tracking is obsolete; brand share of voice and revenue are the real metrics. Brian argues there is no number-one ranking anymore because of personalization and "double JavaScript" AI modes that hide links until clicked. The only metrics that matter are brand share of voice and whether the client is making money.

2. Square Peg SEO means becoming the de facto answer everywhere, not ranking a page. Be the "Jell-O" or "Kleenex" of your category. Search assembles answers from passages, so you force your brand into every answer across AI, images, video, maps, and the SERP. Brian cites a service-contractor client holding 8 of 10 homepage positions across different CMS platforms.

3. The Square Peg system is five pillars. Entity lock-in (same name, services, and descriptors everywhere), distribution over destination (the site converts, everywhere else discovers), co-occurrence engineering (bind your brand to entities that matter), surface area expansion (volume of mentions beats one strong page), and query interception (win the "best way to..." query before Google decides).

4. Build a "Board of Directors" of AI agents in Claude to mass-produce strategy. Brian uploads three skill files plus about twelve markdown agent personas (content director, strategic analyst, PM, SEO/PPC specialist, viral video specialist, neuromarketing psychoanalyst, and more) into a Claude project, with instructions to debate, reach consensus, log to a log.md, and run cascading parallel agent waves. Output is a strategy deck he sells for $2,500 to $3,000.

5. MuVERA produces more comprehensive, less generic AI content. Brian frames MuVERA (Multi-Vector Retrieval Algorithm, Fixed Dimensional Encodings) as the basis of Google's NLP and uses it as a skill to cover a topic from multiple angles (high intent, common, uncommon, rare, fringe) rather than promotional hype, layered with BERT for "360-degree" coverage. (This is Brian's framing; verify the technical characterization before repeating it as fact.)

6. "Link fortification" is tiered link building by another name. Get published on a high-profile third-party site (for example, a search engine journal), then build links to that placement and to your social profiles. Few people build links to their own Twitter or YouTube profiles, which is a missed brand-reinforcement opportunity.

7. In-Page Push ads are unblockable styled-HTML notifications. Unlike older browser push (opt-in rates falling, blocked by iOS, Chrome, and Android), In-Page Push is an HTML element that renders 100% of the time. Dan reports clicks as cheap as half a cent and 10x to 100x cheaper per click than normal paid sources, on a market he states is headed to $35.4 billion in 2026 (his stated figure, source not given).

8. Newsjack Google Trends into AI-built forced-conversion landing pages. Dan exports trending topics from Google Trends as CSV, runs them through Claude agents to pick an angle and write an advertorial "news" page hosted on GitHub Pages or Cloudflare Pages, with menus and footer disabled so the only clickable elements are his offers. Advertorial intermediaries lift conversion from roughly 1 to 3 percent (direct) to about 10 to 13 percent.

9. This ad playbook adapts to local lead gen, not just affiliate. Swap the prompt to produce a state-specific storm or hurricane "preparedness" page for a roofer, with a click-to-call or name/email/phone opt-in wired to a CRM. The same method works for AC, farmers, builders, attorneys (a GLP-1 lawsuit angle), or dietitians.

10. Zone optimization is the single biggest profit lever in push. Dan caps clicks per zone (for example, 100 per zone) to force even traffic distribution and avoid one publisher cluster draining budget. "Zone/source ID" groups publisher sites like zip codes.

Key points

Unidentified speaker (premium PR; name not given)

• Carries roughly $50,000 to $75,000 of PR / news inventory at any given time to post on Yahoo, Globe, AP, and other news partnerships.
• Budget "PBN news network" framing: pay around $100/month for access, or do releases for $20 to $100 each.
• Claim: budget news PBNs get you links and citations but not AI visibility, because the sites are not authoritative enough.
• Hosts note Google deauthorized thousands of Google News sites about two years prior; old pandemic-era Google News sites no longer carry their former authority.
• Brand-perception argument: cheap services devalue a premium brand. He did not want to devalue even his standard $97 press-release package with a $25-tier alternative.

Brian Kato (Square Peg SEO)

Positioning and thesis

• Title: "Square Peg SEO. Make Your Brand Impossible to Ignore." Self-describes as living "at the intersection of cybersecurity and marketing."
• Background: former performing musician, BMI-affiliated rock frontman (played venues from Madison Square Garden to the Viper Room); credits this for his guerrilla-marketing focus.
• Core claim: "There is no number one ranking anymore." Rank trackers are "lying to you" because of personalization and AI mode.
• Google AI mode / AIO is described as "double JavaScript": links do not appear until you click in to reveal them, making rank tracking very hard.
• Search "ranks passages, not pages"; AI extracts a snippet and uses it for citation and assembles answers.
• Be the "Jell-O" or "Kleenex," the de facto brand for the category.
• Client proof point: a service contractor holds 8 of 10 positions on the Google homepage across different platforms (WordPress, Wix, Joomla, HTML, Drupal) and owns the knowledge pack and AI results.

Retrieval mechanics

• BM25 (Best Match 25) is framed as a retrieval layer with three components: retrieval, entity understanding, and answer assembly ("knowledge graph / KP retrieval on steroids").
• "If you're not everywhere, you're not eligible."

The five pillars

1. Entity lock-in. Same name, services, and descriptors everywhere. Reinforce who you are, what you do, where you do it, and what is in it for the consumer.
2. Distribution over destination. Be everywhere (Reddit, social, YouTube, video, in-person events). Site equals conversion; everywhere else equals discovery. Measure calls, leads, and cash, not UTM attribution (calls out iOS 14 breaking meta-ad attribution). Favors physical events such as lunch-and-learns, community goodwill, and philanthropy.
3. Co-occurrence engineering. Combine entities to create node connections (example: "SEO Spring Training" plus "Digital Marketing Conference AZ" / "Chandler, Arizona"). Force entity associations via video and content.
4. Surface area expansion. Volume of mentions beats strength of one page ("1 page = weak, 50 mentions = signal, 500 = dominance"). Build links to social accounts (inspect your Twitter or YouTube profile in Majestic or Ahrefs).
5. Query interception. Win queries like "best way to...", "you are a...", "I need..." before Google decides, via PAAs and "ask anything" tools.

The Claude "Board of Directors" build

• Distribution: QR code to a Google Doc at fvshort.click/seost26 to a Trace / app.tracework screenshot SOP to three downloadable skill files (delivered via Dropbox, flagged "can't virus scan").
• In Claude: Customize, then Skills, then the + (Create/Upload Skill). Upload three files: (1) Fusion Line Gap Analysis (full market gap analysis by website/brand name); (2) the project files, a board of about twelve markdown agents with job descriptions (content director, strategic analyst, project manager, SEO/PPC specialist, viral video specialist, brand strategist, keyword specialist, financial analyst, neuromarketing/psychoanalyst, and more); (3) the MuVERA skill.
• Project instructions: tell the agents they are an advisory board (his is "Future Vine") that must reach consensus on any strategy; "start every conversation by creating a log.md file"; and "invoke cascading waves of agents asynchronously in parallel until the job is completed."
• Live demo: prompt "create a marketing strategy for [website]" or "audit [website]." Example: New England Aquarium (NEAQ), Boston. Output runs in waves (Wave 1 discovery/competitive analysis, Wave 2 strategic foundation) and produces an executive summary, five strategic pillars, 12-month targets/KPIs from board consensus, and persona sections.
• Monetization: he cleans these into a slide deck and charges clients $2,500 to $3,000 each.
• Caveat: "trust but verify... always have that human in the loop."

Geo-grid demo

• Prompt: set up a geo grid and create a keyword list of primary terms and neighborhoods.
• Boston Metro example: 23 official Boston neighborhoods, 80 keywords across 6 clusters, builds a JSON file and a Python script, and outputs a multi-tab spreadsheet (keyword stack, content stack, questions, priority level, search intent, long-tail, page/content format, content type, competition level, notes).

MuVERA and content

• MuVERA framed as Multi-Vector Retrieval Algorithm, Fixed Dimensional Encodings, and as the basis of Google's NLP. Credits JP for introducing him to it.
• BERT framed as bidirectional, "reading with Google."
• MuVERA covers a topic from multiple angles (high intent, common, uncommon, rare, fringe) rather than promotional hype; an HVAC example contrasts Carrier vs Trane warranties. Layer MuVERA plus BERT for "360-degree" coverage.
• Says he killed "80 to 90 percent of our tech stack over the last year"; still uses EAV (Entity, Attribute, Value) tables (a custom GPT) to find keyword gaps via clustering, plus a Guerrilla Marketing GPT for shoestring campaign ideation.

Distribution sprint and "link fortification"

• 48-hour sprint: 5 PR placements, 20 social posts, 10 citations, 1 Reddit thread, 1 YouTube video.
• Distribution map: website, PR sites, GBP, Reddit, YouTube, social, satellites, all reinforcing the same entity relationships.
• "Link fortification" equals tier linking rebranded: publish on a high-profile third-party site (for example, a search engine journal) and build links to that placement.
• Nashville case (with Dre doing most of the work, plus Holly): a high-level subdomain for "best SEO mastermind" got into the AI overview via brand reinforcement and mentions across multiple properties.
• Reddit: works via rage bait and human emotion ("intent behind the intent"); attack from common, uncommon, fringe, and rare angles.
• Closing line: "SEO is no longer just optimization. It's narrative control." / "Make the system adapt to you."

Dan Kurtz (In-Page Push / newsjacking)

Ad type and economics

• In-Page Push (IPP): small notification elements that pop up on a page. Key property is unblockable: it is styled in-page HTML, so ad blockers, iOS, and Android updates do not stop it. Loads 100% of the time as long as the device renders HTML, and avoids the opt-in entirely.
• Push advertising is over a decade old; Dan states it is scheduled to hit a $35.4 billion market in 2026.
• 10x to 100x cheaper per click than normal paid sources; Dan reports half-cent clicks. You run on other people's news sites, not your own.

Networks and setup

• Recommended (tested): Roller Ads (min deposit $50) and MyBid (min deposit $100, even cheaper). Contrast: Outbrain / Taboola charge $3,000 to $15,000 to set up.
• Roller Ads: 4.5 billion impressions/day (his stated figure), min CPC one-tenth of a cent ($0.001).
• His files include ten total ad networks (two tested, eight untested).
• Bidding example: bid $0.005 (half a cent), estimated about 16,000 clicks/day. Some networks permit gray markets (cannabis, gambling) by policy. Start at $5/day and scale.

Landing-page / forced-conversion build

• Do not send cold push traffic direct to an offer; use an advertorial "new style" news-content intermediary page.
• Conversion lift: direct affiliate offer about 1 to 3 percent sale rate; with the advertorial page in between, about 10 to 13 percent. (Cites a Roller Ads case study: pre-landers generate $22 vs $12 for the same link.)
• Page hosted on GitHub Pages or Cloudflare Pages on a purchased domain matching the site name. Pages are not interlinked; each is a standalone landing page.
• The fake news page disables the menu and footer; only the affiliate links / offers are clickable.

Newsjacking workflow (Google Trends to Claude)

• Claude project: "Google Trend Brief Writer."
• Google Trends, then Trending Now, then last 4 hours, then Export CSV; drop into Claude.
• The agent opens the CSV, ranks by volume/percentage, filters out unwanted categories (sports, game scores), picks the top 3, researches, recommends one, and writes the content brief (volume, emotional triggers, push angle, time-to-monetize, audience, monetization ideas).
• Take the brief to the next agent, which writes the article and offers to build a "new style webpage": an artifact with ad spots already defined. Paste in your offers/URLs and publish.
• Models mentioned in passing ("Opus 4.6," "Opus 4.7X," image model "2.5") are loose references, not exact product names.
• Also mentions Feeds / RSS: pipe a tightly filtered trending/local-news RSS feed into a client's site to keep it updated.

Local lead-gen adaptation

• For local, use Google Trends Explore / AI function, set the geo (for example, Florida) and a longer timeframe (past month), pull top/rising queries, pick one, and build an article plus angle.
• Roofing example: a hurricane/storm-preparedness article with a newsjack ad sending to your offer. Drill down by city or zip and refine queries (for example, "asphalt shingles").
• Swap offer type via prompt: click-to-call, or a name/email/phone opt-in wired to a CRM. Works for AC, farmers, builders, attorneys (GLP-1 lawsuit angle), and dietitians.

Campaign ops, tracking, scaling

• Drop your own tracking pixels into the page files; affiliate networks named include Max Bounty and Spring Partner.
• Tracking: DD1 (described as good for affiliate tracking); postback tracking via click ID, conversion fires, result returns. Mentions Microsoft Clarity vs another heatmap tool ("hot dog," a nickname).
• Bidding ladder: start CPC, then Smart CPC, then CPA goal once you have conversions.
• Campaign naming convention: vertical / geo / device / OS / test.
• Zone optimization is the most impactful profitability lever: cap 100 clicks per zone for even distribution and to avoid draining one publisher cluster. Zone/source ID equals clusters of publisher websites, "like zip codes" (credits "Merlino" framing).
• Real campaign (April 8): 19,000 impressions, 125 clicks, 0.66% CTR, 63 cents spent. A separate 3-day run: about 60 cents in clicks returned a $9 commission.
• Targeting controls: geo tiers (tier 1 for English, tiers 2/3 otherwise), devices, OS, browsers, connection type (Wi-Fi only for residential/roofing). Daily budget $5 to $20.
• Scaling: vertical (raise bids), horizontal (new offers/locations), or his favorite, new platforms (max out one network, duplicate the campaign to the next of the ten). Recycle creatives until the story dies.
• Does not trust the networks' automated optimization: "I don't trust it past the end of my fingertips."

Closing sponsor pitch

• A B2B cold-email platform built by Dan and a partner on Instantly, charging $1 per 1,000 emails. Does AI inbox follow-up/nurturing, list cleaning, and warm-up; you only provide the domain. Feature requests submitted as detailed tickets are auto-coded and pushed live the next day. (Product name and URL not stated in the source.)

Slides

Slides (25): Brian Kato, "Square Peg SEO"

Dan Kurtz spoke from slides that were not provided as image files, so no slide gallery is available for his portion (his content is captured in the sections above).

Source

Synthesized from the Day 2 conference recording and the "Brian Kato - Square Peg SEO" deck (25 image-only slides). Dan Kurtz's portion is reconstructed from the recording only, as his slides were not provided as files.