One Questionnaire, Many Libraries: Single-Input Architecture
A questionnaire is not a form. It is a constitutional document — the single source of truth from which twenty-three column prompts draw independently, each reading the same nine sections through its own discipline.
Tommy Saunders
Founder, Windfield Real Estate
April 12, 202613 min read
Direct Answer
What is single-input architecture, and why does one questionnaire drive an entire content operation?
Single-input architecture is the pattern where one structured nine-section questionnaire serves as the sole input for every column prompt in the library system. It is not a prompt — it is an architectural artifact each library reads independently and interprets through its own discipline. The Company Identity library extracts voice. The SEO library extracts keyword clusters. The Sales library extracts subject lines. Same document, multiple simultaneous extractions. Two to fifteen minutes to write; stable across every run of the same campaign.
The most common mistake people make when standing up the library system is not choosing the wrong libraries or misconfiguring the engine. It is treating the questionnaire as a prompt — filling it out like a ChatGPT instruction, keeping it vague because they assume the AI will fill in the gaps. It will not. Unlike a prompt — where vagueness produces a single mediocre output — a vague questionnaire produces twenty-three mediocre outputs, all consistently mediocre in the exact same direction, simultaneously. The system amplifies questionnaire quality. In both directions.
What the questionnaire actually is
A prompt tells a model what to do next. A questionnaire tells twenty-three specialized column prompts what world they are operating in. The distinction matters because the two documents produce fundamentally different kinds of AI behavior.
When you write a prompt, the model reads it once and executes. When you fill a questionnaire, each library reads the entire document and extracts whatever is relevant to its discipline. The Company Identity library reads the brand voice section and derives a register — formal or conversational, technical or accessible. The Sales Enablement library reads the same section and derives a subject-line register — pithy or detailed, exclamation marks on or off, founder name or company name in the from-field. This is why changing one section changes outputs across all libraries — and why getting a section right has compounding returns.
One input, many satellites
The architecture is hub-and-spoke. The questionnaire sits at the center. Specialist libraries hang off it. None of the libraries talk to each other. None of them need to. The questionnaire is the only thing they share, and the only thing they need to share.
Hub-and-spoke · 1 → 5Brief at center
Input
Questionnaire
9 sections · single source
CO-ID
Company Identity
CS
Content Strategy
SEO
Search
SE
Sales
BR-ID
Brand Identity
The annotated questionnaire
The nine sections are not arbitrary. Each one has a specific purpose, a specific set of downstream consumers, and a specific quality signal that separates strong input from weak. Click any section in the panel below to see which libraries read it and what they extract.
Questionnaire — section annotation9 sections active
Click a section ↓
01 Business Type
Entity name + URL — sets namespace for all references
02 Industry
Industry + adjacent categories — drives prose register
03 Company Overview
The strategic spine — one sentence, one claim, falsifiable
04 Goals & Objectives
What success looks like — include decision-making context
05 Target Audience
Primary + secondary audiences and why secondary reads it
06 Brand Voice
Tonal constraints — name what to avoid, not only what to do
07 Competitive Landscape
Named competitors and the structural difference
08 Products & Services
Concrete product names + features — seeds keyword architecture
09 Key Differentiators
The conversion hypothesis — what reader values enough to exchange email
01 · Business Type — how libraries read it
CO-ID
Extracts the entity name for consistent attribution throughout body and bylines. Sets the namespace for every reference.
CS
Derives visual brand anchor — the primary entity whose aesthetic vocabulary informs all strategy concepts.
BR-ID
Creates the brand namespace for all CSS tokens and design-system exports.
SEO
Seeds the entity recognition layer — which name variations to target in structured data.
Quality signal
Include the URL. SEO and Sales use it for canonical links, backlink anchor text, and email footer attribution.
Derives industry framing — technical vocabulary or accessible language, which pain points resonate.
Quality signal
Include sub-categories and adjacent industries. “AI / Prompt Engineering / Content Operations” beats “Technology.”
03 · Company Overview — how libraries read it
CO-ID
Becomes the strategic spine — the argument every section must advance.
BR-ID
Informs the pull quote — the overview often becomes the pull quote directly.
SEO
The primary AEO target — structured as a direct answer to the title’s implied question.
Quality signal
One sentence. One claim. Falsifiable. “Content ops can be AI-native” is weak. “One questionnaire → 23 column prompts → coherent package in <4 min” is strong.
04 · Goals & Objectives — how libraries read it
CO-ID
Sets the assumed knowledge baseline — what to explain, what to state without definition.
SE
Determines the pain point frame — director pain is coordination, founder pain is margin, operator pain is consistency.
SEO
Shapes intent classification — informational vs. navigational vs. commercial query targeting.
Quality signal
Include decision-making context. “Scale ops while maintaining brand coherence across 23 channels” beats “create more content.”
05 · Target Audience — how libraries read it
CO-ID
Determines depth of technical sections — can go deeper knowing technical founders will share.
SEO
Expands the long-tail keyword space without losing topical coherence.
Quality signal
Describe both primary and secondary audiences — and why the secondary reads content not written for them.
06 · Brand Voice — how libraries read it
CO-ID
Sets differentiation language — write against the category, not specific brands.
SEO
Drives keyword gap analysis — terms competitors rank for but haven’t addressed structurally.
SE
Shapes objection handling — Day 3 and Day 8 emails address the implicit “I already use X.”
BR-ID
Informs visual differentiation — anti-pattern against competitor aesthetics.
Quality signal
Include what to avoid. “Never says leverage or unlock” beats “professional and approachable.”
07 · Competitive Landscape — how libraries read it
CS
Generates content strategy directive — topics to own, gaps to exploit, angles to differentiate.
BR-ID
Derives visual differentiation — a system explicitly distinct from competitor aesthetics.
Quality signal
Name competitors and articulate the structural difference — “they do X, we do Y, here’s why it matters.”
08 · Products & Services — how libraries read it
SEO
Seeds the full keyword architecture — 12 to 18 long-tail variants from primary names.
CO-ID
Becomes the semantic spine — which concepts get definitional treatment vs. assumed knowledge.
Quality signal
List specific product names and features. “23 Column Prompts · Notion Template System” gives libraries concrete terms.
09 · Key Differentiators — how libraries read it
SE
Defines the conversion hypothesis — what the reader values enough to exchange an email.
Quality signal
Describe the reader’s mental state at the capture form — “read 70% of the article, understands the concept” produces a much better sequence.
The interpretation matrix
The same section — read simultaneously by four different libraries — produces four structurally different extractions. This is not redundancy. It is the mechanism by which one document generates coherent content across twenty-three column prompts. The example below shows the Brand Voice section as read by each library.
Brand Voice sectionSame section · 4 extractions
Section 06 value“Systems-minded. Precise. Operator-first. No hype. Never says ‘leverage’ or ‘unlock.’ Against single-prompt tools. Differentiated by orchestration and structural coherence. They solve generation. We solve coordination.”
Library
What it extracts
What it produces
Company Identity
Extracts the structural argument frame: “They solve generation. We solve coordination.” This becomes the central differentiation claim — cited in the lede, elaborated in body, closed in the conclusion.
“The question is not whether AI can generate content — it clearly can. The question is whether it can coordinate.”
SEO & Web
Extracts keyword gap opportunity: competitors rank for “AI writing tool” but weakly for “content orchestration” and “multi-agent content system” — high-intent, lower competition.
Extracts the objection to handle: “I already use Jasper.” Day 3 addresses this as a category distinction — “Jasper writes for you. We coordinate for you. Different job.”
Day 3: “You probably already have a writing tool. That’s not the problem.”
Brand Identity
Extracts visual anti-pattern: competitors use bright blue-on-white, gradient-heavy, rounded aesthetics. The brand system is explicitly dark, editorial, precise — positioned in a different category.
Tokens: dark page, IBM Plex Serif display. Anti-pattern: no gradients on white.
Notice that none of these libraries coordinated. The Company Identity library did not tell Sales what objection to handle. Brand Identity did not ask SEO what competitors look like. Coherence emerges from the shared input, not from inter-library communication. This is the architectural guarantee — and it is only possible because every library reads the same document.
A prompt has one reader. A questionnaire has twenty-three. That difference is the whole architecture.
Architecture Series
Stability and reuse
A questionnaire is stable across runs. If the campaign is the same, the questionnaire does not change between Monday’s run and Friday’s run. This is what makes the pipeline reliable: re-running with the same input produces the same coordinated package — same argument, same tokens, same voice — with whatever incremental data the libraries pulled at runtime.
Reuse is why questionnaire authorship matters more than prompt authorship. A great prompt produces one good output. A great questionnaire produces twenty-three coordinated outputs every time it runs, for the lifetime of the campaign. The leverage is permanent.
9
Sections
23
Column prompts
~5m
Author time
Why vagueness compounds
Every library fills questionnaire gaps with its own interpretation. The Company Identity library guesses one thing, Sales another, SEO a third. By the time the package ships, the three are subtly arguing past each other. Vague input, expensive output.
Frequently Asked Questions
5 Questions
What is single-input architecture?+
A pattern where one structured document — the questionnaire — feeds many specialist libraries simultaneously, each extracting different fields through its own discipline. Same nine sections, coherent output across all column prompts.
How is a questionnaire different from a prompt?+
A prompt is one instruction to one model. A questionnaire is a structured document many models read independently. Same document, multiple simultaneous extractions. The questionnaire is architectural; the prompt is operational.
What are the nine sections?+
Business Type, Industry, Company Overview, Goals, Target Audience, Brand Voice, Competitive Landscape, Products & Services, Key Differentiators. Every section is consumed by multiple libraries — Voice by Company Identity and Brand Identity and Sales, Industry by Company Identity and Sales, and so on.
How long does the questionnaire take to fill?+
Two to fifteen minutes for an experienced author. It is the rate-limiting step in the pipeline — once written, every column prompt runs automatically and the package ships in roughly two minutes of compute.
What happens if a section is vague?+
Every library fills the gap with its own interpretation. The result is twenty-three outputs that coherently disagree. The Company Identity library guesses one thing, Sales another, SEO a third — and the package ships arguing past itself.
About the author
Tommy Saunders
Founder, Windfield Real Estate
Building the AI-native content operations system for business operators who need predictable output, not AI experiments.