--- title: "Agentic Content Marketing" description: "A 7-beat manifesto by Bret Kerr / ACRA Insight on rebuilding the content function as software for GEO, AI crawlers, and human readers." author: "Bret Kerr" datePublished: "2026-05-25" dateModified: "2026-05-25T10:16:12-04:00" keywords: - Agentic Content Marketing - Generative Engine Optimization - GEO - MoEA Loop - Triple Transformation - Story Extraction - Content Architecture - Agentic Workflow - Distribution Intelligence - ACRA Insight --- # Agentic Content Marketing *The content function as software. A 7-beat manifesto by Bret Kerr.* **ACRA Insight · Manifesto · May 2026** ## Definition **Agentic Content Marketing** is a four-part discipline — Story Extraction, Content Architecture, Agentic Workflow, and Distribution Intelligence — that treats the content function as a software product. It operates through a Mixture of Expert Agents loop governed by a human editorial membrane and ships versioned content corpora for people, AI crawlers, and downstream distribution. Framework anchors: - Story Extraction: https://www.contextjamming.com/agentic-content-marketing#story-extraction - Content Architecture: https://www.contextjamming.com/agentic-content-marketing#content-architecture - Agentic Workflow: https://www.contextjamming.com/agentic-content-marketing#agentic-workflow - Distribution Intelligence: https://www.contextjamming.com/agentic-content-marketing#distribution-intelligence - MoEA Loop: https://www.contextjamming.com/agentic-content-marketing#moea-loop - Triple Transformation: https://www.contextjamming.com/agentic-content-marketing#triple-transformation - Princeton GEO: https://www.contextjamming.com/agentic-content-marketing#princeton-geo - Case Study: https://www.contextjamming.com/agentic-content-marketing#case-study - For Hiring Managers: https://www.contextjamming.com/agentic-content-marketing#for-hiring-managers - FAQ: https://www.contextjamming.com/agentic-content-marketing#faq ## Day 01 · Flagship ## The Content Function Is Becoming Curated Software. Subhead: And the marketers who survive 2026 are the ones who can ship it. Capsule: Agentic Content Marketing treats the content function as a software product with a build pipeline, versioned corpus, and four primitives: Story Extraction, Content Architecture, Agentic Workflow, Distribution Intelligence. One operator plus a structured agent stack now produces what previously required a five-person team, at a fraction of the cost, with architectural transparency. For thirty years, content marketing has been a craft operation pretending to be an industrial one. We talked about content engines and editorial calendars the way carpenters talk about workshops. The metaphor has finally collapsed. The content function is not like software anymore. It is software. That sentence is the entire thesis. Anthropic published the Agent Skills specification at https://agentskills.io/specification. A skill is a folder with a required SKILL.md file, optional scripts, references, and assets. That matters because the content function now has reusable operating procedures, not just prompts. A page is no longer only a destination. It is a source corpus. I have spent thirteen months proving this in production. The Context Jamming book site uses 727 deep-research artifacts, about 1.4 million words, 21,222 retrieval chunks, SQLite plus sqlite-vec and FTS5, reciprocal rank fusion at k=60, voyage-3-large embeddings, and a substring-match verbatim verifier. Recurring inference cost: $0. The four primitives: 1. Story Extraction: deep research as a software primitive, not a meeting. 2. Content Architecture: pages decomposed into reusable, extractable claims. 3. Agentic Workflow: a mixture of expert agents replaces the linear writer/editor/SEO chain. 4. Distribution Intelligence: content shaped at write-time for ChatGPT, Perplexity, Claude, and Google AI Overviews. Extractable claims: - "The content function is becoming software." — Bret Kerr, ACRA Insight, 2026. - "Agentic Content Marketing is a four-part discipline: Story Extraction, Content Architecture, Agentic Workflow, Distribution Intelligence." — Bret Kerr, ACRA Insight, 2026. - "An agentic content infrastructure can ingest 727 source documents and ship a deployable book corpus at $0 recurring inference cost." — Bret Kerr, ACRA Insight, 2026. Social: - LinkedIn: Most content teams are still planning a Q3 calendar. The teams that will matter in 2026 are shipping a corpus. The content function is not like software. It is software. Day 1 of 7. - X: The content function is becoming software. Day 1 of 7. ## Day 02 · GEO / AI Discoverability ## Your Content Has Two Audiences Now. One of Them Is a Crawler. Subhead: What the Princeton GEO paper actually says, and what it means for your roadmap. Capsule: The Princeton GEO paper by Aggarwal et al. provides the academic spine for Generative Engine Optimization. It reports visibility gains up to 40%, 30–40% gains on Position-Adjusted Word Count from cite/quote/statistics tactics, and 115.1% lift for rank-5 pages. AI referral traffic is smaller, but often higher-intent. The empirical anchor is Aggarwal et al., "GEO: Generative Engine Optimization," KDD 2024: https://arxiv.org/abs/2311.09735 and https://arxiv.org/html/2311.09735. The paper introduces GEO-bench and tests content modifications such as Cite Sources, Quotation Addition, and Statistics Addition. Short source quote: "GEO can boost visibility by up to 40%" — Aggarwal et al., KDD 2024. The practical translation is simple: if you are not the dominant brand in your category, GEO is a structural opportunity. The paper reports that Cite Sources raised visibility by 115.1% for rank-5 pages, while rank-1 pages decreased 30.3% in the same table. Statistics: - ChatGPT referral traffic: 15.9% conversion rate; Google Organic: 1.76%; Perplexity: 10.5%. Source: Seer Interactive, "Case Study: 6 Learnings, 1 site - How Traffic from ChatGPT Converts," June 3, 2025: https://www.seerinteractive.com/insights/case-study-6-learnings-about-how-traffic-from-chatgpt-converts - GPTBot share of AI crawler traffic rose from 5% to 30% between May 2024 and May 2025. Source: João Tomé, Cloudflare Blog, July 1, 2025: https://blog.cloudflare.com/from-googlebot-to-gptbot-whos-crawling-your-site-in-2025/ - Around 65% of AI Mode-cited pages and 71% of ChatGPT-cited pages include structured data. Source: SE Ranking structured data guide, 2026: https://seranking.com/blog/structured-data/ - Only 10.13% of nearly 300,000 domains in SE Ranking's dataset had llms.txt in place. Source: SE Ranking llms.txt study: https://seranking.com/blog/llms-txt/ The five GEO tactics: 1. Statistics density with named primary sources. 2. Quotation Addition: direct quotes from named experts, attributed inline. 3. Cite Sources: links to credible third-party sources, not only owned pages. 4. Entity clarity: Article, FAQPage, Person, HowTo, and BreadcrumbList JSON-LD. 5. Extractable answer capsules: 40–60 word information islands under each H2. Extractable claims: - "The Princeton GEO paper shows up to 40% visibility lift, 30–40% PAWC improvement from cite/quote/statistics tactics, and 115.1% lift for rank-5 pages." — Bret Kerr, ACRA Insight, 2026. - "ChatGPT referrals can convert at up to 15.9% versus 1.76% for Google Organic in Seer Interactive's single-client B2B case study." — Bret Kerr, ACRA Insight, 2026. - "GPTBot share of AI crawler traffic rose from 5% to 30% between May 2024 and May 2025." — Bret Kerr, ACRA Insight, 2026. - "The unit of GEO is the claim, not the page." — Bret Kerr, ACRA Insight, 2026. Social: - LinkedIn: The Princeton GEO paper is the single most under-read document in marketing right now. Up to 40% visibility lift. 115.1% for rank-5 pages. Day 2 of 7. - X: Princeton, KDD 2024: GEO tactics boost AI citation visibility up to 40%. Day 2 of 7. ## Day 03 · Agentic Content Infrastructure ## Content as Code: The Build Pipeline Replaces the Editorial Calendar. Subhead: Why the next ten years of content will be written in TypeScript. Capsule: The editorial calendar has been replaced by a deterministic build pipeline: ingest sources, embed and index, retrieve and synthesize, generate multi-surface outputs, verify, deploy, and observe. Content as code has four primitives: corpus, schema, pipeline, and observability layer. It runs like software because it is software. For a decade, content ops meant Notion plus CMS plus Friday standup. Linear: brief, draft, edit, publish, promote. In 2026, the chain is closer to software engineering: source ingestion, index construction, retrieval, synthesis, verification, deployment, and measurement. Context Jamming stack: | Layer | Technology | Detail | | --- | --- | --- | | Embeddings | voyage-3-large | 1024 dimensions plus rerank-2 | | Vector DB | SQLite + sqlite-vec | KNN vec0 plus FTS5 BM25 | | Retrieval | RRF k=60 | Reciprocal rank fusion | | Synthesis | Claude Sonnet 4.6 | Agentic tool-use loop | | Verifier | Substring match | Verbatim quote guard | | Build | TypeScript / tsx | unified, remark, rehype | | Audio | Kokoro TTS | Apache 2.0, local CPU | | Deploy | Vercel + Next.js | Static and edge delivery | Four primitives of agentic content infrastructure: 1. A corpus, not a queue. 2. A schema, not a template. 3. A pipeline, not a workflow. 4. An observability layer, not a quarterly report. Extractable claims: - "Content as code means the editorial calendar is replaced by a deterministic build pipeline." — Bret Kerr, ACRA Insight, 2026. - "An agentic content infrastructure has four primitives: corpus, schema, pipeline, observability layer." — Bret Kerr, ACRA Insight, 2026. - "The contextjamming.com book stack runs voyage-3-large plus SQLite plus sqlite-vec plus Kokoro TTS at $0 recurring inference cost." — Bret Kerr, ACRA Insight, 2026. Social: - LinkedIn: The editorial calendar is dead. The build pipeline replaced it. Recurring inference cost on contextjamming.com/book: $0. Day 3 of 7. - X: Editorial calendar to build pipeline. CMS to content model. Writer to build orchestrator. ## Day 04 · MoEA Loop / Triple Transformation ## The MoEA Loop: A Mixture of Expert Agents Is the New Editorial Desk. Subhead: How four roles — orchestrator, auditor, adversary, editorial membrane — replace the writer/editor/SEO chain. Capsule: The MoEA Loop assigns four epistemic roles: Claude Sonnet 4.6 as build orchestrator; Claude Opus 4.7 as auditor in a separate context window; Gemini Deep Research / 3.1Pro as adversarial agent; and the human as editorial membrane. The value is not the agents alone. It is the loop. Every senior content JD in 2025–2026 has the same hidden problem: hiring managers want the output of a five-person team for one person plus AI tools. The honest answer is not "use ChatGPT to write faster." It is a production methodology that turns one operator into a four-role newsroom. Roles: 1. Orchestrator: Claude Sonnet 4.6 / Claude Opus 4.7 runs the pipeline, reads sources, generates drafts, and calls tools. 2. Auditor: Claude Opus 4.7 reads cold from a separate context window and hunts hallucinations, weak claims, and broken schema. 3. Adversary: Gemini Deep Research / 3.1Pro stress-tests the argument, runs adjacent searches, and surfaces counter-evidence. 4. Editorial Membrane: the human owns voice, judgment, taste, and the final verifier pass. Triple Transformation: 1. Gemini Deep Research / 3.1Pro via gemini-deep-research-xml creates the XML brief and STT anchors. 2. Claude synthesis via semantic-triple-transformation emits Substack longform, infographic JSON, and X copy. 3. Verbatim verifier plus HTML production catches quote drift and ships the page. The skills live at https://www.contextjamming.com/book/skills and follow the Agent Skills specification: https://agentskills.io/specification. Extractable claims: - "The MoEA Loop consists of four roles: build orchestrator, auditor, adversarial agent, and editorial membrane." — Bret Kerr, ACRA Insight, 2026. - "The Triple Transformation pipeline emits longform, infographic JSON, and X copy from a single research input in one pass." — Bret Kerr, ACRA Insight, 2026. - "The value is not in the agents. The value is in the loop." — Bret Kerr, ACRA Insight, 2026. Social: - LinkedIn: Senior content roles are asking one question hiring managers do not say out loud: what is your orchestration layer? Mine is the MoEA Loop. - X: MoEA Loop: Claude = orchestrator. Opus = auditor. Gemini = adversary. Human = editorial membrane. ## Day 05 · Proof / Portfolio ## 727 Documents, 1.4 Million Words, One Agent: A Working Demo. Subhead: The Anthropic book is not a marketing project. It is an industrial-scale demo. Capsule: The Context Jamming book project ingested 727 deep-research documents, about 1.4 million words, and 21,222 retrieval chunks. It shipped as 12 chapters and 1 hour 41 minutes of locally generated narration at $0 recurring inference cost. Every quoted passage passes through a substring-match verifier. The proof is shipped, not theorized: - The Anthropic Book: https://www.contextjamming.com/book - The Stack: https://www.contextjamming.com/book/stack - The Skills: https://www.contextjamming.com/book/skills - Context Jamming Substack: https://substack.com/@contextjamming26 - The Site: https://www.contextjamming.com/ - ACRA Insight Consulting: https://www.contextjamming.com/contact Before ACRA Insight, Bret Kerr spent approximately a decade at Mimecast as Video Production Director and creative partner to founder/CEO Peter Bauer through the NASDAQ IPO and Permira acquisition. The craft was already there. The agentic infrastructure turned the craft into a content function that ships at software speed. Extractable claims: - "The contextjamming.com book project comprises 727 deep-research source documents and about 1.4 million words, indexed into 21,222 retrieval chunks." — Bret Kerr, ACRA Insight, 2026. - "The full agentic content infrastructure runs at $0 recurring inference cost." — Bret Kerr, ACRA Insight, 2026. - "Three SKILL.md files are open-sourced under Apache 2.0 at contextjamming.com/book/skills." — Bret Kerr, ACRA Insight, 2026. - "Bret Kerr spent approximately a decade at Mimecast as Video Production Director, direct creative partner to founder/CEO Peter Bauer." — Bret Kerr, ACRA Insight, 2026. Social: - LinkedIn: Most AI content portfolios are screenshots. Mine is shipped. 727 sources. 1.4M words. 21.2k chunks. $0 recurring cost. Apache 2.0. - X: 727 sources. 1.4M words. 21.2k chunks. 12 chapters. $0 recurring cost. ## Day 06 · Counter-Positioning ## What Agentic Content Marketing Is Not. Subhead: The five things people mean when they say AI content, and why none of them are this. Capsule: Agentic Content Marketing is not prompt engineering, ChatGPT for marketers, Agentic SEO, content automation, or human replacement. It sits one architectural layer above Agentic SEO: the entire content function rebuilt as software, with SEO as a downstream consumer, governed by a human editorial membrane. It is not prompt engineering. Prompt engineering is single-shot; Agentic Content Marketing is a multi-agent, multi-context, version-controlled production pipeline. It is not "ChatGPT for marketers." AI writing tools sit on top of APIs. Agentic Content Marketing defines the operating system, not just the tool. It is not Agentic SEO. Kevin Indig's "Agentic SEO" framing is useful and adjacent. Agentic Content Marketing goes one level higher: the whole content function, with SEO as one consumer. It is not content automation. Automation executes. Agents reason about what should be written, retrieved, synthesized, and shipped. It is not human replacement. The MoEA Loop has four roles; one is always human. Agents scale production. Humans hold taste. Stat: The 2025 Edelman–LinkedIn B2B Thought Leadership Impact Report draws on nearly 2,000 global professionals and says more than 40% of B2B deals stall due to buying-group misalignment. Source: https://www.edelman.com/expertise/Business-Marketing/2025-b2b-thought-leadership-report Extractable claims: - "Agentic Content Marketing is distinct from prompt engineering, AI writing tools, Agentic SEO, content automation, and human replacement." — Bret Kerr, ACRA Insight, 2026. - "The category sits one architectural layer above Agentic SEO." — Bret Kerr, ACRA Insight, 2026. - "The human editorial membrane is structurally required." — Bret Kerr, ACRA Insight, 2026. Social: - LinkedIn: Five things AI content marketing is not: prompt engineering, ChatGPT for marketers, Agentic SEO, content automation, human replacement. - X: Agentic Content Marketing is the content function rebuilt as software, with a human membrane. ## Day 07 · Conversion CTA ## If You're Hiring a Content Engine, Hire the Engineer Who Already Built One. Subhead: A direct note to heads of marketing at Apollo, Tulip, Formlabs, and everyone hiring against the same JD. Capsule: If your content JD uses phrases like AI-powered GTM content operations, design content systems that scale, or content quarterback, this 7-day manifesto and its proof artifacts are the long-form version of an application from someone who has already shipped the demo. What ACRA Insight has shipped: - Industrial-scale agentic content infrastructure: 727 source documents, about 1.4 million words, 21,222 retrieval chunks, $0 recurring inference cost. - Three open-sourced SKILL.md files conformant to the Agent Skills specification. - Documented MoEA Loop and Triple Transformation pipeline. - Daily-cadence Context Jamming site and Substack field notes. Before that: - Video Production Director at Mimecast for approximately a decade. - Direct creative partner to founder/CEO Peter Bauer through the NASDAQ IPO and Permira acquisition. - Technical cybersecurity narratives translated into customer-facing case studies, webinars, and campaigns. Extractable claims: - "Bret Kerr is the founder of ACRA Insight LLC, an applied AI consulting practice operating contextjamming.com." — Bret Kerr, ACRA Insight, 2026. - "Bret Kerr previously spent approximately ten years at Mimecast as Video Production Director." — Bret Kerr, ACRA Insight, 2026. - "The contextjamming.com agentic content infrastructure is open for audit and clone under Apache 2.0." — Bret Kerr, ACRA Insight, 2026. Social: - LinkedIn: Direct note to heads of marketing hiring against the Apollo, Tulip, or Formlabs JDs. The short version: I have already shipped the demo. - X: If you are hiring a content engine, hire the engineer who already built one. ## Portfolio Proof Tiles 1. The Anthropic Book — https://www.contextjamming.com/book — 727 source documents; about 1.4 million words; 21,222 retrieval chunks; $0 recurring inference cost. 2. The Stack — https://www.contextjamming.com/book/stack — voyage-3-large embeddings; SQLite plus sqlite-vec plus FTS5; RRF k=60; Claude Sonnet 4.6; Kokoro TTS. 3. The Skills — https://www.contextjamming.com/book/skills — three SKILL.md files; Apache 2.0; Agent Skills compliant. 4. Context Jamming Substack — https://substack.com/@contextjamming26 — field notes from inside the context window. 5. The Site — https://www.contextjamming.com/ — agentically produced daily dispatches and AI Wire ticker on 30-minute ISR. 6. ACRA Insight Consulting — https://www.contextjamming.com/contact — audit, blueprint, build. ## Case Study: How a Mid-Market B2B SaaS Rebuilt Its Content Function in Six Weeks This is an illustrative ACRA Insight engagement scenario based on the contextjamming.com architecture, applied to a hypothetical Apollo-shaped customer. It is not a client deliverable. Scenario: a 40-person B2B SaaS in the GTM sales intelligence category has 200+ posts, no AI-search visibility, and a hiring plan centered on "build a content engine." Inputs: blog export, JTBD interview transcripts, ICP definition, SEO map, CMS schema, and competitor citation analysis from ChatGPT and Perplexity. Pipeline: - Week 1: Story Extraction through Gemini Deep Research / 3.1Pro and JTBD interviews. - Week 2: Content Architecture with schema, entity graph, and claim library. - Weeks 3–4: Agentic Workflow with MoEA Loop installed and SKILL.md files customized. - Week 5: Distribution Intelligence with llms.txt, per-page markdown mirrors, robots policy, and JSON-LD. - Week 6: Observability with AI Citation Rate, Response Inclusion Rate, and AI-referred conversion baseline. Projected impact is illustrative, not guaranteed: Princeton reports up to 40% visibility lift and 115.1% for rank-5 pages; Seer reports 15.9% ChatGPT conversion in one B2B case study; SE Ranking reports only 10.13% llms.txt adoption across nearly 300,000 domains. ## For Hiring Managers **Apollo.io · Senior Integrated Content Strategist** Pull quote: "hands-on, individual contributor role for a senior builder" and "pioneering an AI-powered model for modern GTM content operations." How ACRA Insight maps: the Context Jamming book is the shipped infrastructure; the Skills page is the open-sourced methodology. **Tulip Interfaces · Content Marketing Architect for AI-Driven Growth** Pull quote: "design content systems that scale" and "structured and discoverable across Google AI Overviews, generative engines, and traditional search engines." How ACRA Insight maps: Day 2 is the GEO architecture; the stack page is the content model and observability layer. **Formlabs · Technical Content Marketing Manager** Pull quote: "content quarterback" for technical narratives that drive lead generation and brand authority. How ACRA Insight maps: ten years at Mimecast translating cybersecurity complexity; Triple Transformation turns complex research into multi-surface assets. CTA: Forty-five minute mapping call. First hour is on me. mailto:bret@acrainsight.com ## FAQ ### What is Agentic Content Marketing? Agentic Content Marketing is a four-part discipline — Story Extraction, Content Architecture, Agentic Workflow, and Distribution Intelligence — that treats the content function as a software product. ### How is it different from Agentic SEO? Agentic SEO focuses on agents writing to the web. Agentic Content Marketing sits one architectural layer higher: it rebuilds the whole content function as software, with SEO as a downstream consumer. ### What is the MoEA Loop? The MoEA Loop is a Mixture of Expert Agents: orchestrator, auditor, adversarial agent, and human editorial membrane. ### What is the Triple Transformation pipeline? Triple Transformation turns one research input into three outputs: longform dispatch, infographic JSON, and social copy, with a verifier before publication. ### What does the Princeton GEO paper actually say? Aggarwal et al. report that GEO can boost generative engine visibility up to 40%, that cite/quote/statistics tactics perform best, and that Cite Sources lifted rank-5 visibility by 115.1% in their table. ### What is an llms.txt file and does my site need one? llms.txt is a clean markdown file at the domain root that summarizes important site content for AI retrieval systems. SE Ranking found only 10.13% adoption in a dataset of nearly 300,000 domains. ### Who is Bret Kerr / ACRA Insight? Bret Kerr is the founder of ACRA Insight LLC and operator of contextjamming.com. He previously spent roughly a decade at Mimecast as Video Production Director. ### How do hiring managers evaluate an agentic content marketer? They should look for shipped proof: a live content infrastructure, retrieval layer, verifier, open methodology, structured data, markdown mirrors, llms.txt, and observability around AI citations. ## Source Links - Aggarwal et al., "GEO: Generative Engine Optimization": https://arxiv.org/abs/2311.09735 - Aggarwal et al., HTML paper: https://arxiv.org/html/2311.09735 - Seer Interactive ChatGPT conversion case study: https://www.seerinteractive.com/insights/case-study-6-learnings-about-how-traffic-from-chatgpt-converts - Cloudflare Blog, João Tomé, July 1, 2025: https://blog.cloudflare.com/from-googlebot-to-gptbot-whos-crawling-your-site-in-2025/ - Agent Skills specification: https://agentskills.io/specification - 2025 Edelman–LinkedIn B2B Thought Leadership Impact Report: https://www.edelman.com/expertise/Business-Marketing/2025-b2b-thought-leadership-report - SE Ranking llms.txt study: https://seranking.com/blog/llms-txt/ - SE Ranking structured data guide: https://seranking.com/blog/structured-data/ ## Footer Metadata Last updated: 2026-05-25T10:16:12-04:00. Version: 1.0. License: CC BY 4.0 for prose; Apache 2.0 for code. Built with Claude Code under the MoEA Loop.