The Memo Spec

How We Build Content That AI Models Cite

Context Memo uses a versioned content framework — the Memo Spec — to produce reference-grade articles structured for both human decision-makers and AI search engines. Every memo generated by the platform follows these principles.

Current: v4.411 major versions since January 2026

Core Principles

These apply to every memo, regardless of type or version.

Dual Audience

Every memo is written for two readers simultaneously: a human buyer evaluating solutions, and an AI model deciding what to cite. The structure serves both.

Analyst, Not Advocate

Memos read like analyst briefings — factual, balanced, and grounded in verifiable sources. Brands are positioned within a landscape, not marketed above it. This is what makes the content citable.

Verifiable Claims Only

No fabricated statistics, testimonials, or customer names. Every claim must be traceable to provided context. "Not publicly disclosed" is always preferred over a guess.

Structure as Signal

Heading hierarchy, section ordering, and content density are deliberate signals — not formatting choices. They're designed to match how AI models extract and rank information.

Versioned and Iterable

The spec is versioned like software. Every memo is tagged with the spec version that produced it, so we can measure what works and improve systematically.

Honest Limitations

Memos include transparent assessments of where a solution may not be the right fit. This builds credibility with both human readers and AI models that value balanced sources.

Memo Types

Different buyer questions require different content structures. Each type is purpose-built for a specific stage of the decision journey.

Comparison

Mid-Funnel

Side-by-side analysis of two solutions. Includes a quick-reference comparison table, individual breakdowns, and key differences — written for buyers actively evaluating options.

Alternative

Mid-Funnel

Multi-option evaluation for buyers searching for alternatives to a specific vendor. Each option gets a focused analysis with a clear "best for" recommendation.

Industry Brief

Top-of-Funnel

Vertical-specific guide addressing the pain points, compliance requirements, and use cases relevant to a particular industry.

How-To Guide

Top-of-Funnel

Educational, step-by-step content that teaches readers how to accomplish something. At least 70% educational — vendor positioning comes at the end, not the beginning.

Gap Analysis

Mid-Funnel

Answers a specific buyer question that isn't well-covered in existing content. Provides a direct answer, problem context, and tool comparison.

Product Insight

Bottom-of-Funnel

Coverage of product updates and new capabilities. Frames changes in terms of what buyers care about: what changed, why it matters, how to evaluate it.

Synthesis

Mid-Funnel

Comprehensive deep-dive that synthesizes insights from multiple sources into a single definitive reference. The longest format — designed to be the one article AI models cite instead of any individual source.

Citation Response

Mid-Funnel

Strategic content that covers the same ground as pages AI models currently cite — but from the brand's perspective, with proprietary expertise and data woven throughout.

Version History

The Memo Spec evolves based on what we learn about AI citation patterns. Each version represents a meaningful improvement to how content is structured, not just cosmetic changes.

4.4

v4.4Editorial Standard

May 28, 2026

A universal quality floor applied to every memo, regardless of brand. Bans the most common AI-generated tells — opening clichés, filler verbs, generic adjectives, sales-pitch closers — so brand voice shapes the prose without slipping into formulaic SaaS-blog patterns.

  • Catches the prose patterns AI models penalize as low-signal
  • Bans clichés like "in today's rapidly evolving...", "delve into", "cutting-edge", "in conclusion"
  • No specific dollar amounts in pricing — link to vendor pricing pages instead (pricing changes; stale numbers hurt credibility)
  • Structural rules prevent empty FAQs and "contact vendor for pricing" comparison tables
  • Brand voice can refine the standard, never lower it
4.3

v4.3Brand Voice Fingerprint

May 28, 2026

Memos now sound like the brand actually wrote them. The system extracts a brand's writing fingerprint — sentence rhythm, vocabulary, recurring patterns, opening style — from their existing published work, then injects that fingerprint as binding editorial guidance into every memo.

  • Extracted directly from the brand's own writing — not inferred from a persona or archetype
  • Captures rhythm, vocabulary, taboo phrases, and example openings the brand actually uses
  • Applies across every memo type and every creation path
  • Graceful fallback for brands without an extracted fingerprint
4.2

v4.2Separated Brand Positioning

May 13, 2026

Synthesis and citation-response memos now keep the brand entirely out of objective tool lists, then give it a dedicated "How [Brand] Fits Into This Landscape" section below for honest differentiation. The split makes the rest of the content more citable.

  • Brand removed from neutral roundups, comparison tables, and ranked lists
  • Dedicated fit section: positioning, grounded differentiators, an honest "not for" statement
  • Required acknowledgment of what the brand does not replace
  • Applies to the formats AI models cite most heavily
4.1

v4.1Cited-Content Grounding

April 29, 2026

Memos that should reference real, fetchable sources now actually do. Comparison, how-to, and industry memos pull full content from the URLs AI models currently cite and use that material as grounding. A grounding gate prevents these formats from publishing without it.

  • Cited-source fetching extended to comparison, how-to, and industry memos
  • Pre-generation gate refuses to publish ungrounded content when grounding was attempted but failed
  • Source URLs persisted on every memo with cited content
  • Brands without scan data are unaffected — gate only fires when grounding was attempted
4.0

v4.0Citation Authority

April 16, 2026

A philosophical shift from brand-promotional to objective reference content. Memos now position the brand as one credible option within comprehensive market analysis — matching the patterns AI models actually cite. The content earns its citation by being useful, not by promoting.

  • Brand analyzed alongside competitors with the same lens, not promoted above
  • Synthesis memos rewritten as analyst-grade market reports
  • Gap-fill memos lead with the direct answer, then cover competing approaches objectively
  • Third-party evidence emphasis: review data, market stats, pricing structure with links to vendor pricing pages (no dollar amounts that go stale)
3.7

v3.7Data-Driven Structure

March – April 2026

Memos now adapt their structure based on competitive intelligence. Before generating content, the system analyzes what top-cited content in each category actually looks like — word count, heading patterns, data density — and sets structural targets accordingly.

  • Competitive structural analysis before generation
  • Automated quality gates prevent incomplete content from publishing
  • Funnel-stage-aware content strategy
3.5

v3.5Modular Architecture

March 2026

Shifted from monolithic templates to a modular block system. Each memo assembles from a shared foundation plus type-specific blocks, ensuring consistency across all content while allowing each format to do what it does best.

  • Shared content principles applied uniformly
  • Type-specific sections assembled dynamically
  • Entity recognition for AI knowledge graphs
3.0

v3.0Citation Research

February 2026

A major overhaul informed by research into what content AI models actually pick up and cite. Introduced structural patterns designed specifically for AI extraction — not just SEO, but how large language models identify and surface authoritative content.

  • Executive summary sections for quick AI extraction
  • Question-format headings that match how users prompt AI
  • Multi-source synthesis with 3,000+ word deep dives
  • Strict external sourcing standards
2.5

v2.5Commercial Layer

February 2026

Added the layer that converts citation traffic into pipeline. Memos can now surface relevant CTAs and brand offers. Introduced two new content types: gap-fill memos that answer specific buyer questions, and product-deploy memos covering new capability launches.

  • "Next Step" sections with brand-specific offers and CTAs
  • Gap-fill and product-deploy memo types added
  • Citation-response: strategic variations of content AI models currently cite
  • Stricter source rules: no brand self-sourcing, real external URLs only
2.0

v2.0Analyst Voice

February 2026

Established the editorial identity: every memo reads like a third-party analyst briefing, not a marketing page. Content is written for the buyer — the person evaluating solutions — with the brand positioned as one factor in a broader landscape.

  • Third-person analyst perspective
  • Buyer-persona-targeted writing
  • Expert voice insights with verified attribution
1.0

v1.0Foundation

January 2026

The original memo format. Factual, structured content designed to be referenced by AI assistants. Four content types covering the core buyer journey.

  • Comparison, industry, how-to, and alternative formats
  • Configurable brand voice and tone
  • Source-cited reference structure

Want this for your brand?

Context Memo generates AI-optimized reference content using the Memo Spec. Every article is structured to be cited by ChatGPT, Claude, Perplexity, and Gemini.

Request Access