Strategic Foundation
Brand work is the highest-leverage discipline in business-to-business (B2B) SaaS — and the most misunderstood. Same conviction-pillars-voice block onboards a new employee and onboards every AI agent in your stack.
The framework — strategy first
THE FOUNDATION-BEFORE-AMPLIFICATION RULE
"AI helps amplification. It does not inherently generate impact. If what you offer and why is fuzzy, AI multiplies the fuzziness. If the target market isn't clear, AI multiplies the blast radius. If proof is weak, AI multiplies claims without substance. So before you multiply with AI, ask: what are you multiplying?"
The discipline this work exists to enforce. Brand work has shifted from "the asset that informs the website" to "the substrate every agent, every prompt, and every AI-generated output is built on." An AI tool running on a sloppy Brief produces sloppy outputs faster, more broadly, and across more surfaces than any human team ever could. Get the foundation wrong and you scale wrongness.
The three critical pieces of the foundation are (1) what you offer and your value, (2) target market, and (3) customer receipts. Brand & Positioning is where you write (1). Ideal Customer Profile (ICP) & Audience is where you sharpen (2). Reviews & Social Proof and Customer Marketing carry (3). All three have to be strong before AI amplification — the discipline of which lives in AI Operating Model — is worth turning on.
There's a second reason this matters more, not less: the buyer's research path now runs through AI tools that cite third-party authority over your owned content. In every CMO conversation I've had in 2025–2026, the same pattern: most B2B buyers now use LLMs in their research phase, and most complete most of the journey before contacting sales. To be the brand the AI shortlists, you need a defensible category position that earned-media authority sources will pick up and amplify. Weak positioning doesn't get cited. What "weak" looks like in the wild: the same company described as "the simplest scalable cloud," "an under-the-radar company providing cloud services," and "public cloud platform powered by 17 data centers" across BusinessWire, Motley Fool, and SiliconANGLE in the same news cycle. Three different value statements, one company, every AI tool reading them gets a smeared signal. The fix isn't more PR — it's tighter foundation work upstream.
Brand work is the highest-leverage discipline in B2B SaaS — and the most misunderstood. The misunderstanding: most CMOs still treat brand as marketing's marketing doc — an asset for the website, the deck, the launch. That framing under-prices brand by an order of magnitude in the AI era. The same conviction-pillars-voice block that onboards a new employee onboards every AI agent in your stack. It is the context layer for all three of your audiences — customers, humans inside your company, and the machines that now produce work on your behalf.
This work teaches the discipline of writing that context layer down. The output is not a deck. The output is the first six sections of your Operator Brief, populated — and from those six sections every downstream prompt and agent inherits the voice, pillars, and positioning that make your output sound like one company.
by the end of this section you will have written five artifacts in your own words: your core conviction, four brand pillars, voice rules, positioning statement, and the differentiator you'd defend last.
Before AI, the brand-and-positioning doc had two audiences — customers and the people inside the company building work for them. Now there's a third: the AI agents producing work alongside your team. The doc you write here serves all three. The same conviction-pillars-voice block grounds your customer-facing positioning, briefs a new hire on their Day-30 message, and grounds the system prompt that keeps every AI output sounding like you and not like a generic model.
THREE AUDIENCES, ONE SOURCE OF TRUTH
Customers — Sections 1.9 (conviction), 1.14 (positioning statement), and 1.15 (differentiator) drive your website, sales deck, and category narrative.
Employees — Section 2.10 (brand pillars) plus 1.11–1.13 (voice DOs, DON'Ts, forbidden language) plus 1.14 (positioning) onboard a new hire on Day 1. If they can write a credible Day-30 LinkedIn post from these alone, the doc is doing its job.
Agents — Sections 1.9 through 1.15, transformed into a system-prompt context block. Every downstream agent in this book — the Web Operations, Performance Marketing, and Field Marketing Agents in AI Operating Model, plus the Content / Email / LinkedIn / Customer-Stories / PR agents in Content & SEO / Email & Lifecycle / LinkedIn & Social / Customer Marketing / PR & Comms — imports from this single source. The Sections 1.1–1.8 context (company, category, stage, ACV, motion, vertical, regulatory posture, buyer profile) frames every prompt that follows.
Notice that the agents and the employees need the same content. The litmus test for whether your brand doc is operationally sharp is the same one it has always been — could a smart new hire produce a credible Day-30 message from this alone? If yes, your AI can do the same with one prompt. If no, your AI will hallucinate or hedge, exactly the way an under-briefed new hire would. The onboarding test is the AI test.
THE BRAND-AS-MOAT THESIS — A ONE-PAGE ARGUMENT FOR YOUR CEO
Claim 1 — AI commoditizes models, not context. When every competitor has access to the same Claude, GPT, and Gemini, the only thing they don't have is your brand doc. Your conviction, pillars, voice, and forbidden language are the inputs nobody else can copy, because they're built from your conviction, not pulled from training data.
Claim 2 — Sloppy positioning compounds. Pre-AI, a vague brand meant a few off-tone emails per week. Post-AI, a vague brand means every one of your agents is broadcasting drift at scale, 24/7. The penalty for unclear brand went from linear to exponential.
Claim 3 — Brand investment now has 2× ROI. Same doc, two workforces — humans and agents — both grounded in it. The cost stays the same; the leverage doubles. The CMO who maintains the cleanest brand doc has the most defensible AI advantage in their market.
Bottom line — Cutting the brand budget now cuts the context layer your AI inherits. Generic in, generic out, at scale. The CEO ask: protect the discipline that lets AI sound like you and not like a model.
The 1-of-2 thesis — why positioning consistency is now structural
LLMs converge. When a buyer asks ChatGPT, Perplexity, or Claude "what are the best tools for [your category]?" the answer typically names one or two brands. The third spot is fragile. Beyond three, citation rate collapses. The narrowing happens because LLMs optimize for consensus across sources, not authority from any single one — and the corpus refreshes constantly.
The operating implication for brand: positioning that is described identically across every surface (your homepage, your G2 profile, your LinkedIn company page, your newswire releases, every analyst briefing, every podcast appearance, every founder LinkedIn post) gives an LLM more consensus signal to ingest. The brand that owns the same one-sentence positioning across 15 places ends up on the 1-or-2 shortlist. The brand that has six versions of the same positioning across 15 places gets read as noise and falls off.
The senior-operator habit: the conviction, positioning statement, and category claim that lives in your Operator Brief is the single sentence that gets repeated verbatim on every surface, by every speaker, in every press release, every podcast intro, every founder post, every G2 profile description. Not paraphrased. Not adapted to the channel. The discipline isn't creative — it's repetition. This is the operating premise behind every prompt in the rest of the playbook.
Most B2B SaaS marketing teams operate without a written brand. Their pillars live in someone's head. Their voice is enforced by editorial vibes rather than rules. Their positioning is a deck someone made for a board meeting two years ago. The result — every time a new piece of content goes through review, three people argue about tone. Every time a prospect lands on your site after seeing a LinkedIn ad, the page reads like it was written by a different company. Every time a rep tries to differentiate from a competitor in a deal, they reach for a different talking point. Now multiply each of those by the number of AI agents in your stack and you see the AI-era version of the problem.
Brand drift is not a creative problem. It is an operating problem. Without a written brand, you cannot scale messaging, you cannot onboard a new content writer in under three weeks, and your agents will produce output that sounds like a generic model — because they have nothing to anchor on.
OPERATOR WISDOM — THE BRAND AUDIT
Take the homepage hero, the most recent LinkedIn ad, the latest customer email, and a page of the most recent investor update. Read them back-to-back out loud. If they don't sound like the same company, you have brand drift. The fix is not a rebrand — it is writing the brand down.
THE UPSTREAM CHECK — DON'T WRITE CONVICTION FIRST
The most common Brand & Positioning failure: a CMO sits down to write conviction, pillars, and positioning before doing the upstream work that defensible brand is built on. What comes out is a doc that sounds confident in the conference room and falls apart in front of a buyer who asks "but why you over [competitor]?" Brand without competitive understanding and a named right-to-win is a brochure. The conviction has to survive a competitor's funding round; you can only know that if you've already mapped the competitors.
Three upstream inputs feed every artifact here. Do them — even lightly — before you write the conviction. The depth lives in adjacent sections of the playbook, but the synthesis lives in Section 2.16 of your Brief.
You cannot write a defensible conviction without naming the alternatives the buyer is choosing between. The full work — battlecards, win/loss themes, competitive positioning matrices — lives in Product Marketing & Competitive Intel. For Brand & Positioning, you need the lightweight version: the top 3 named competitors, the one-sentence reason each one wins deals you lose, and the one-sentence stance your brand takes that they can't credibly copy. If you can't fill that in for all three, you're not ready to write conviction — you're ready to talk to ten customers.
Identify the market where you have the strongest right to win with your value proposition. The goal is not the widest blast radius — it is unmistakable relevance. Right-to-win is the intersection of three things: a real buyer pain you uniquely solve, a segment where you have or can build genuine credibility, and a place where the competitive gap is on your side. It is the answer to "if we had to win one segment, which one would survive a funding-round-sized punch from the category leader?"
HOW TO FIND IT — LOOK FOR WHERE THE EASY WINS ARE
Two signals point at your right-to-win segment. Run both in parallel.
| SALES SIGNAL — "WE LOVE LEADS LIKE THIS" | PRODUCT-LED GROWTH (PLG) SIGNAL — "THEY GOT VALUE FAST" |
|---|---|
| Which leads do sales get most excited about? Which companies or buyers have no tradeoffs in the conversation? Who has the fastest buying cycles? |
Who activates fastest after sign-up? Who expands naturally without a CSM nudge? Who has the lowest churn? |
Then validate with customer conversations. Find the common traits. That's your right-to-win.
Proof this works. Teams that lock in their ICP commonly see ~2× paid marketing efficiency. What focus enabled: cut nearly half the event roadmap, tightened paid campaign targets, sharpened their PLG motion, shut down non-critical campaigns. Demand went up across the board. The discipline isn't "say no to non-ICP work" — it's "say yes to the segment where the math actually works."
Before you write conviction, talk to five customers. Not for testimonials — for the upstream language that conviction is built from. The five questions below are the canonical script:
Listen for the recurring phrase — the one three or four customers use unprompted to describe what you do. That phrase is almost always closer to your real positioning than anything your team has written. It's also the bridge between this work and Reviews & Social Proof (Reviews) + Customer Marketing where customer evidence compounds.
The output of the three upstream inputs is a single sentence captured in Brief Section 6.16. The sentence has three parts: (1) the segment where you have the strongest right to win, (2) the named competitors you win against in that segment, (3) the structural reason you win that they cannot easily close. One sentence. The brand work below — conviction, pillars, voice, positioning, differentiator — is the artistic expression of that one sentence. Without the sentence, the brand work is improvised.
The synthesis of competitive landscape (Product Marketing & Competitive Intel), the easy-wins analysis above, and your five-customer conversations. One sentence that the brand work below is built on. Saves to Brief Section 6.16 — every downstream prompt + agent reads from this.
Saved to Brief Section 6.16. If you can't write this sentence yet, work Product Marketing & Competitive Intel (competitors) and the five-customer conversations above before returning to write conviction.
Brand identity has three components: a Core Conviction (the belief you operate from), Brand Pillars (the four-to-five recurring beliefs your work expresses), and a Brand Voice (how those beliefs sound in language). Each is a distinct artifact. Each goes into your Operator Brief.
Core Conviction.
A core conviction is a one-sentence belief about the world, not about your product. It is contrarian enough to be interesting, true enough to act on every day. "We make great software" is not a conviction — it is a brochure. "B2B buyers are tired of being marketed at; they want to be respected" is a conviction. "Frontline workers deserve the same software experience executives get" is a conviction. Your conviction explains why your company exists in a way that survives a pricing cut, a feature war, and a competitor's funding round.
Write three convictions. Pick the one that makes you most uncomfortable to say out loud. Comfort is the enemy of differentiation.
Brand Pillars.
Pillars are the four-to-five recurring beliefs your work expresses. They are the load-bearing posts under your conviction. They show up in product decisions, content topics, hiring, and customer interactions. Pillars are not features. They are not values. They are the consistent stances your company takes when no one is watching. Most B2B SaaS companies need exactly four pillars. More than five and they dilute, fewer than three and they look thin. Below is the framework template. Populate it as you read.
THREE REAL-WORLD PILLAR EXAMPLES TO CALIBRATE AGAINST
Linear (project tracking) — Speed of execution · Quality of craft · Built for builders · Default to async. Notice: each is a stance, not a feature. "Default to async" is a contrarian operating belief that changes how the product is built.
Stripe (payments) — Increase the GDP of the internet · Developer-first · Quality at scale · API-first. The first pillar is an existential mission, not a feature claim. The second through fourth are stances against the legacy of payments software.
Beacon Frontline (the running example throughout) — Worker-first · Day-of-shift · Operations not comms · Multi-unit native. Each one rejects a competitor stance. "Operations not comms" actively distances Beacon from chat-app competitors — that's how a pillar earns its place.
Brand Voice.
Voice is the language of your brand. It has four dimensions — tone, language, perspective, style — and it requires both a "what we are" column and a "what we are not" column. The negation is what makes the voice operationally useful. "Confident" without "not arrogant" is too vague to enforce. The voice rules below should be specific enough that a freelance writer could pass a blind test after reading them.
OPERATOR WISDOM — THE FREELANCE TEST
Hand your Brand Voice rules and three brand pillars to a freelance copywriter you've never worked with. Have them write a 200-word product blurb. If it sounds like you, your rules are sharp. If it doesn't, your rules are too abstract. The fix is more specificity — pull from real sentences your team has written that you'd put on the homepage.
CAPTURE IN YOUR OPERATOR BRIEF — Section 2.9 / Section 2.10 / Sections 1.11–1.13
Once you've worked through this section, capture: your Core Conviction in Brief Section 6.9, your four pillars (comma-separated) in Brief Section 6.10, your Voice DOs in Section 2.11, Voice DON'Ts in Section 2.12, and Forbidden Language in Section 2.13. Sign and date.
The framework template — populate as you go. Every field below saves to your Operator Brief automatically; you can also view it at /brief.html anytime.
Three artifacts. Conviction, pillars, voice. Type your own version in the fields below — autosaves as you type, fills into every prompt on the site.
| Dimension | What we are | What we are not |
|---|---|---|
| Tone | ||
| Language | ||
| Perspective | ||
| Style |
Saved to Brief Section 6.9 (conviction), Section 2.10 (pillars), Sections 1.11–1.13 (voice DOs, DON'Ts, forbidden language). See it all at your Brief.
The fill-in-the-blanks positioning statement and the category frame you operate in. The output is one defensible sentence.
Saved to Brief Section 6.1 (company), Section 2.2 (category), Section 2.14 (positioning statement), Section 2.15 (differentiator). The composed positioning statement is generated automatically from your fields.
Three levels. Category narrative at the top, solution theme in the middle, proof points at the bottom. A writer should be able to draft the homepage hero from Level 1, the product page from Level 2, and case study quotes from Level 3.
Saved to your Brief. Available to every prompt on the site as [MESSAGE HIERARCHY].
Three traps appear in pillar work. The first is the "values trap" — the team writes their HR-deck values (Excellence, Customers First, Integrity) and calls them pillars. Values are personal commitments; pillars are public stances. Customers don't buy from companies that promise integrity. They buy from companies that promise a specific belief about their world.
The second is the "feature trap" — the team writes pillars that describe what the product does (AI-Powered, Mobile-First, Secure by Design). Features change; pillars shouldn't. If your pillar gets invalidated when you ship a roadmap update, it isn't a pillar.
The third is the "too many trap." Five is a healthy ceiling. Six pillars means you actually have two pillars and four pet topics in disguise. Cut down to five.
Most B2B SaaS companies position around features and lose. The reason is structural — features are commodities once a competitor copies them, and they always get copied. Category positioning is durable because it claims a frame, not an attribute. "The CRM" is a category position. "The CRM with the best email integration" is a feature claim. The first survives a competitor's email-integration release. The second does not.
Category positioning has three moves available — define a new category (hardest, highest-reward), redefine an existing category (most common), or own an existing category (only available to the dominant player). Most growth-stage B2B SaaS companies should attempt the second. The OG framework template below walks you through each move.
CAPTURE IN YOUR OPERATOR BRIEF — SECTION 1.11 / 1.12
Once you've worked through positioning: capture your one-sentence positioning statement in Brief Section 6.14, and the single differentiator you'd defend last in Brief Section 6.15. The differentiator is the load-bearing post — if you removed it, the deals you'd lose first.
Most teams build messaging hierarchies upside-down. They start with the proof points ("we have 200 enterprise customers"), then work up to the solution theme, and finally to the category narrative — and the result is messaging that talks about the company, not the buyer's world. The hierarchy works the other way. Level 1 is the category narrative — the big claim about how the buyer's world is changing. Level 2 is your solution theme — the mechanic by which you address it. Level 3 is your proof — the evidence that the mechanic works. When you hand a writer the hierarchy, they should be able to write the homepage hero from Level 1, the product page from Level 2, and the case study quotes from Level 3 — top-down, in that order.
BRAND & POSITIONING AND OPERATOR BRIEF SECTION 2 ARE THE SAME ARTIFACT, FRAMED TWO WAYS
This work is the teaching. Operator Brief Section 6 is the form. The output of this work is Brief Sections 6.1 through 2.16, populated and signed. Don't treat them as separate deliverables — the Brief is the doc, this work teaches the doc. Once Section 2 is populated, every prompt and agent downstream from this point in the playbook imports from it.
TWO DIFFERENT JOBS, TWO DIFFERENT OUTPUTS
Most teams use “brand” as one word and end up with vague work. The discipline starts with naming the two halves separately:
| BRAND STRATEGY | BRAND MESSAGING |
|---|---|
| What it is. The long-term plan for the brand — conviction, pillars, audience, competitive stance, visual identity, voice. The internal operating system. | What it is. The external expression of the strategy — tagline, headlines, ad slogans, sales pitches, the specific words on the page. |
| What it answers. Who are we, who do we serve, what do we stand for, how do we show up consistently across every surface? | What it answers. Given the strategy, what do we say to this buyer, on this channel, at this moment in their journey? |
| How often it changes. Rarely. The strategy should hold for 18–36 months at a minimum. | How often it changes. Often. Messaging should be tested, refined, and rotated quarterly. |
| Where it lives. Brief Sections 6.9 through 2.13 — conviction, pillars, voice DOs/DON’Ts, forbidden language, positioning. | Where it lives. Brief Sections 6.14 through 2.16 — positioning, differentiator, right-to-win — plus the messaging hierarchy above and every downstream prompt. |
The two are inseparable in practice. A clear strategy with weak messaging gets ignored. Strong messaging without a strategy gets contradicted by the next campaign. The Brief captures both; this work teaches both; the downstream channels execute against both.
PATIENCE BEATS EXECUTION SPEED IN THE FIRST 90 DAYS
Brand work is iterative. Teams that try to ship a brand identity in a single creative brief almost always have to redo it 12 months later when the work doesn’t hold up. The three-phase ideation funnel below is the discipline. Each phase has a different job; each phase ends with a decision point that locks in inputs for the next.
| PHASE | WHAT YOU DO | DECISION POINT TO CLOSE THE PHASE |
|---|---|---|
| 1. Discovery | Brainstorm wide. Gather every stakeholder. Pull every adjacent reference, competitor, parallel-industry inspiration. Encourage bad ideas — bad ideas surface the constraints. Document everything; judge nothing. | A pool of 30–80 raw ideas, themes, references, and audience insights captured in one place. |
| 2. Direction | Narrow. Pick the philosophy. Eliminate ideas that don’t align with the conviction, the audience, or the business goal. Cluster what remains into 2–3 distinct directional bets. | A signed directional brief naming the philosophy and the 2–3 directional bets that will be tested in refinement. |
| 3. Refinement | Iterate. Test each directional bet against real audience segments — sales reps, customers, ICP-fit prospects. Refine the wording, the visual identity, the tone, the proof points until one direction is unmistakably stronger than the others. | A locked brand strategy and messaging system that holds up under feedback and ships to every downstream prompt in the playbook. |
Three rules that hold across the phases. Stay focused on conviction at every gate — ideas that conflict with the conviction get cut before they consume calendar time. Gather feedback before locking — the strategy that ships is the strategy that survives 5 customer conversations, not the strategy the team likes most. Celebrate small wins through the iteration — the work feels slow; naming the wins keeps the team in the work long enough to finish it.
A PILLAR IS WHAT YOU CLAIM. AN EXCLUSION IS WHAT YOU REFUSE.
Every brand has a parallel document the team rarely writes down: the list of messages, emotions, and features the brand will not use, evoke, or claim. Without it, well-meaning team members ship copy that drifts away from the brand. With it, the team has a fast filter for every borderline campaign.
| EXCLUSION TYPE | WHAT TO NAME | EXAMPLE |
|---|---|---|
| Messages to avoid | Themes, claims, or topics the brand will not engage with — even when they would be commercially convenient. | “We do not make fear-based security claims. We do not endorse polarizing political content. We do not benchmark against competitors by name in paid media.” |
| Emotions to avoid | Feelings the brand will not evoke in its audience — even when those feelings would drive short-term conversion. | “We do not evoke FOMO. We do not evoke shame. We do not evoke distrust of the buyer’s own team.” |
| Features to avoid | Product capabilities the brand will not lead with — even when the engineering team is proud of them. | “We do not lead with AI-as-feature. We do not lead with integrations as the headline. We do not over-claim performance metrics we can’t reproduce in the customer’s environment.” |
| Tonal traps to avoid | Voice patterns the brand will not use — even when the writer is in a hurry. | “We do not use jargon (‘synergy’, ‘leverage’, ‘empower’). We do not use exclamation points outside customer quotes. We do not use stock photography for hero imagery.” |
Exclusions belong in Brief Section 6.13 alongside forbidden language. The two together — what we’ll never say plus what we’ll never write — give every downstream agent a clean filter before it ships work in your voice.
EVERY BRAND CONCEPT IS A TWO-PART CLAIM
A complete brand concept names why the audience needs the solution in their language, and then names the proof that lets them believe you. Most brand work delivers half of this — usually the need framing — and leaves the buyer to assemble the proof themselves.
The proof half is where most B2B brands underinvest. The pattern below is the menu of proof types every brand should be able to point to. The strongest brand concepts use at least three of these in combination; weak concepts rely on only one.
| PROOF TYPE | WHAT IT IS | WHEN IT EARNS THE MOST TRUST |
|---|---|---|
| Case studies and success stories | Named customer outcomes — what the customer faced, what changed after, the quantified result. | Mid-to-late stage in the buyer journey. The single highest-converting brand asset for considered B2B purchases. |
| Customer testimonials | First-person quotes from named customers in their own words. | Throughout the funnel. Most effective when the testimonial customer holds the same buying-committee role as the prospect. |
| Industry certifications and awards | Third-party recognition — ISO, SOC 2, HIPAA, G2 leader badges, analyst awards. | Risk-averse buyers (regulated industries, enterprise security reviews, healthcare, financial services). |
| Data and statistics | Concrete numbers from your customer base — reduction percentages, time-saved, ARR growth attributed. | CFO-grade buyers. The proof type that wins board-level conversations. |
| Expert endorsements | Named industry experts, analysts, or influencers who vouch for the solution. | Category-creation moments and early-market positioning when the buyer can’t yet evaluate from peer evidence. |
| Transparency and accountability | Public roadmaps, open-source contributions, published incident reports, honest disclosure of what the product does not do. | Buyers who’ve been burned by overpromising vendors before. The trust premium compounds over time. |
The discipline isn’t to collect every proof type — it’s to know which two or three your buying committee actually weighs, and to make those unmistakably visible everywhere the brand shows up. The persona deep-dive work in ICP & Audience tells you which proof types matter most to each buying-committee role.
HOW THIS SECTION CONNECTS TO THE PROMPT PACK BELOW
Once Brief Sections 6.1–2.16 are populated, the six prompts in the Prompt Pack do real work — they produce a positioning statement that already uses your conviction, persona messaging that already references your pillars, a competitor differentiation grid that already names your real differentiator, brand exclusions captured for every downstream agent to filter against, and a reusable system-prompt block (Prompt 6) that hands all of the above to every downstream agent in this playbook. The prompts are the execution layer. The framework is the IP.
The prompt pack
Each prompt is a named, named-by-what-it-does deliverable. Click any card to expand the paste-able body. Run against your Operator Brief.
Five copy-paste prompts. Open ChatGPT, Claude, or Gemini. Paste a prompt. Run it. The output of one prompt feeds into the next.
READ THIS ONCE BEFORE ANY PROMPT IN THIS BOOK
These prompts assume you've populated your Operator Brief (the worksheet that lives in /Operator-Brief-Worksheet.docx). When a prompt asks for OPERATOR BRIEF, paste the relevant Brief sections rather than typing context from scratch.
Your output then arrives in your voice, against your buyers, using your differentiators. Not [BRACKETED] generics. The Brief is the difference between an LLM helper and a tool that sounds like you.
Prompt 1
A one-sentence positioning statement (in your voice, using your pillars) plus three variants you can A/B test in messaging.
Prompt 2
A messaging matrix: one row per buying-committee persona × four columns (success metric, primary objection, headline, proof asset). Headlines arrive in your voice; proof assets reference your real pillars.
Prompt 3
A 1-page brand voice rulebook with 5 do's, 5 don'ts, and two sample paragraphs (good vs. off-brand).
Prompt 4
A 250-word category narrative we can post to LinkedIn or use as the opening of an investor deck.
Prompt 5
A 5-row grid (us vs 3 competitors) on dimensions that actually matter to the buyer — not vendor-defined feature checklists.
Prompt 6
A reusable Markdown context block, formatted for system prompts, that imports your brand doc into any AI agent. Drop into Claude Project instructions, custom GPT instructions, or any agent's grounding doc. Run this once after Brief Section 6 is populated; re-run whenever Section 2 (ICP) changes. This is the prompt every downstream agent depends on.
The agent spec
How to install this agent
Maintains the brand voice rules, positioning statement, and brand pillars as the single sentence repeated verbatim across every surface (the 1-of-2 LLM thesis demands consensus through repetition, not paraphrase). Refreshes the voice forbidden-language list. Pressure-tests positioning against market signal. Audits cross-surface consistency — does every G2 profile, LinkedIn page, podcast intro, and press release say the same sentence the same way? The strategic owner the Brand Voice Agent enforces against.
| Task | Frequency | Duration | Output goes to |
|---|---|---|---|
| Weekly Brand Voice Agent score scan | Weekly Mon 10:00 | ~20 min | Head of Brand |
| Weekly Win/Loss + Market Watch scan | Weekly Mon 10:30 | ~30 min | Head of Brand + PMM |
| Monthly Brand Drift digest | Monthly 1st | ~90 min compile | Head of Brand + VP Marketing |
| Monthly forbidden-language update | Monthly 15th | ~30 min | Brand Voice Agent owner |
| Monthly positioning consistency audit | Monthly 20th | ~60 min | Head of Brand + PMM |
| Quarterly full refresh | Quarterly Q-1 days | ~4 hours | VP Marketing + CEO if positioning changes |
| Annual category narrative review | Annually | ~6 hours | CEO + VP Marketing + Board if material |
Scheduled (cron-style):
| Schedule | What it runs |
|---|---|
0 10 * * 1 | Weekly drift scans |
0 9 1 * * | Monthly Brand Drift digest |
0 9 15 * * | Monthly forbidden-language update |
0 9 14 1,4,7,10 * | Quarterly full refresh |
Event-driven:
| Event | What it runs |
|---|---|
| Market Watch flags category-shifting competitor move | Pressure-test category narrative within 7 days |
| Brand Voice Agent calibration agreement drops below 90% | Audit rubric for rule drift within 48 hours |
| Win/Loss surfaces a positioning-contradiction theme | Pressure-test affected pillar within 14 days |
| CEO or Board flags positioning concern | Run ad-hoc category narrative review within 14 days |
| Source | Type | Cadence | Required? |
|---|---|---|---|
| Operator Brief (Sections 6, 8) | Markdown | Read every run | Required — THE artifact under management |
| Brand Voice Agent score history | Postgres | Weekly | Required |
| Win/Loss Agent themes (brand / positioning / voice tags) | Markdown | Per-interview | Required |
| Market Watch competitor positioning snapshots | Markdown | Daily | Required |
| Customer Marketing reference language patterns | Markdown | Weekly | Required |
| Brand voice rules + positioning statement (current) | Markdown (Git) | Read every run | Required |
| Forbidden-language list (current) | YAML | Versioned | Required |
| Output | Format | Target path | Audience |
|---|---|---|---|
| Monthly Brand Drift digest | Markdown | /brand/drift/YYYY-MM.md | VP Marketing + Head of Brand |
| Forbidden-language list updates | YAML diff + Git PR | /brand/forbidden-language.yaml | Brand Voice Agent |
| Monthly positioning consistency audit | Markdown | /brand/consistency/YYYY-MM.md | Head of Brand |
| Quarterly refresh proposals | Markdown + Git PR against Brief Sections 6 + 8 | /brand/quarterly-refresh/Q<n>.md | VP Marketing + Brief Sync Agent |
| Annual category narrative review | Markdown | /brand/category-narrative/YYYY.md | CEO + VP Marketing + Board |
| Voice-calibration playbook (updates) | Markdown | /brand/calibration-playbook.md | Brand Voice Agent owner + drafting agent owners |
| Trigger condition | Escalate to | Within |
|---|---|---|
| Quarterly refresh missed deadline | VP Marketing + Head of Brand | Immediate |
| Brand Voice Agent agreement < 90% for 2+ weeks | VP Marketing + Head of Brand | Same week |
| 3+ Win/Loss themes contradict same positioning pillar in a month | VP Marketing + CEO | < 14 days |
| Competitor category-shift requires positioning response | VP Marketing + CEO | < 7 days |
| Annual category narrative review surfaces material category misfit | CEO + VP Marketing + Board | Same week |
| Platform / tool | Used for | Required? |
|---|---|---|
| Claude Project + Git | Brand artifact versioning + reasoning surface | Required |
| Postgres | Drift signal history | Required |
| GitHub / GitLab API | PRs against the Brief | Required |
| Slack API | Monthly digest delivery + drift alerts | Required |
Evals — output quality checks:
Hallucination defense — specific checkpoints:
First-run checklist — 5 steps from spec to running agent: