Brand Strategy
The founder's funnel, from strategy to coverage
Buyers used to find you through a search engine. Now they ask an assistant. A founder in Jakarta types a question into ChatGPT. A procurement lead in Singapore asks Claude who the serious vendors are. A reporter checks Perplexity before a call. In each case a model answers with a short list of names, and your job is to be one of them.
That is harder than it sounds, and it is where most PR goes wrong. Founders reach for press before they can say, in one clean line, what they do and who they do it for. They send pitches a reporter cannot place and a model cannot verify. The coverage does not come, and when it does it is vague, so the next reader stays confused.
This guide is for founders and lean teams who want press and have not yet done the brand work underneath it. The argument is simple. An unplaceable company is unquotable, to a reporter and to a model alike. Get the brand right first, which means positioning, messaging, and voice, and every pitch, every page, and every AI answer gets easier and more consistent. We will build that foundation in order, then show how it turns into coverage and citations. Nothing here needs a big budget. It needs clarity, and the discipline to fix it before you spend a single pitch.
Chapter 1
Picture a technology reporter at Tech in Asia on a normal morning. Two hundred unread emails. A deadline before lunch. She opens a pitch that begins, "We are an AI-powered platform that empowers businesses to unlock growth." She cannot tell what you sell, who buys it, or why it matters this week. She does not reply. She does not save it for later. She deletes it, and she files the sender under noise.
This is what pitching before the brand is clear looks like from the other side of the inbox. Press is not step one. It multiplies a foundation that either exists or does not. When the foundation is fuzzy, you multiply fuzz, and you do it in front of the exact people you most wanted to reach.
A reporter's job is to tell readers, in a sentence, why a thing deserves their attention. If she cannot write that sentence from your pitch, there is no story for her to write. She will not invent your positioning for you. She has twenty clearer pitches stacked behind yours, and she will reward the ones that did the thinking first.
So the first test is blunt. Can a stranger, thirty seconds into your home page, say out loud what you do and who it is for? If that sentence wobbles, everything downstream inherits the wobble: the pitch, the landing page, the founder bio, the quote you give on a call.
There is now a second reader who is just as literal, and far less forgiving. When someone asks ChatGPT, Claude, Perplexity, or Gemini to name vendors in your category, the model assembles an answer from what it can find and verify. It is looking for a claim that is specific, repeated in the same shape across several places, and checkable against a source. A company that describes itself three different ways on three different pages gives the model nothing stable to hold. So the model reaches for a competitor who said one clear thing, one clear way, everywhere.
The two readers want the same thing. A human reporter needs a line she can put in a headline. A model needs a line it can retrieve and stand behind. Fuzzy brand fails both at once. This is why the old sequence, get press first and the positioning will follow, no longer works, if it ever did. The model does not wait for your positioning to settle. It answers today, with whatever is consistent today.
The cost is not abstract. Start with the pitch. You get one clean shot at a given reporter on a given story. Send her something she cannot place and you have not merely failed to land coverage. You have spent the introduction. Re-pitching the same person the same muddled idea next month makes you look worse, because now you are the founder who still cannot explain the company.
Then there is the money you spend to look busy. Founders who have skipped the brand work tend to buy activity: a wire release no desk reads, a run of thin blog posts, a logo in a pay-to-play list. A release that opens with a leading provider of innovative solutions is filler with a dateline, and the editors who see it learn to skip the next one from you. None of it compounds, because none of it says the same sharp thing twice. You end up with motion and no position.
There is a slow cost inside the company too. When the team cannot agree on one line, sales decks, job posts, and investor updates all drift. Each person fills the gap with a private guess. By the time a reporter or a model samples three of those surfaces, they meet three different companies wearing the same logo.
The companies that get covered are placeable. A reporter can set them on a shelf: this is a payroll tool for restaurants, this is a cross-border payments company for Southeast Asian exporters, this is a cybersecurity firm for hospitals. Placeable companies are quotable, because the reporter already knows which story they belong in and which desk should own them. Quotable companies, said the same way often enough, become the names a model repeats.
Notice the order. Placeable comes first. Quotable follows. Cited follows that. You cannot skip to the end by pitching harder. A louder pitch for an unplaceable company is only louder fuzz.
This is the case for doing the brand work before the outreach. It is not a decorative exercise, no mood boards required. It is the practical business of becoming describable: one audience, one difference, one line, said consistently everywhere a human or a machine might look. The next three chapters build that, in order. Positioning decides who you are for and what you do differently. Messaging turns that into a few provable claims. Voice makes it sound like you. Do them first, and the pitch almost writes itself.
Takeaway: If a reporter cannot place you in one line and a model cannot verify you against a consistent claim, no amount of pitching will land. Become describable first.
Chapter 2
Positioning answers two questions and holds them together: who you serve, and what you genuinely do differently for them. The real audience, not the one you wish you had. The real difference, provable today, not a roadmap promise. Get these two right and most other decisions fall out of them for free.
Founders resist narrowing because it feels like shrinking the market. The instinct is to say "we help businesses," because every business could, in theory, buy. But "we help businesses" is unplaceable. It names no desk, no reporter, no shelf. Compare two descriptions of the same company. One: an HR platform for companies. The other: payroll built for restaurants with hourly and tipped staff. The second is smaller on paper and far stronger in practice. A restaurant owner reads it and thinks, that is for me. A reporter reads it and knows which story it belongs in. A model reads it and has a clean, narrow claim to attach to your name.
Narrow does not cap your growth. It gives you a beachhead you can win and a sentence people repeat for you. You can widen later from a position of strength. You cannot widen from a position of fog. The companies that own a category almost always started by owning a corner of it.
Here is the test to run before anything else. Fill in one sentence, out loud, with no filler: we help a specific who do a specific what by a specific how, unlike the status quo. Then read it to someone outside your company and watch their face. If they nod and can repeat it back, you have a position. If they ask so what does it actually do, you have a slogan, and slogans do not survive a reporter's inbox.
Good one-liners are concrete and almost boring. Cross-border payroll for Asian startups hiring remote engineers. Fraud detection for Southeast Asian e-commerce checkouts. Carbon accounting for food manufacturers. Each names a buyer and a job. Each tells you where it belongs. None of them reaches for revolutionary, next-generation, or AI-powered, because those words describe nothing and a reporter has read them a thousand times this year.
This is the part founders miss. Positioning is not only internal clarity. It routes you to a specific desk and a specific reporter, and that routing is most of whether a pitch lands.
Newsrooms are organized by beat. A fintech reporter at The Business Times covers payments, lending, and digital banks. A general tech reporter at e27 or Tech in Asia covers funding and launches across the board. A deals reporter at DealStreetAsia tracks rounds and the companies behind them. If your position is sharp, you know within seconds which of these people is yours. Fraud detection for e-commerce checkouts belongs to the fintech beat, and you can name the three reporters who have covered payments fraud in the last year. If your position is fuzzy, you cannot route at all, so you blast everyone, and a blast to everyone reads as spam to each one.
The sharper the position, the shorter and more obvious the target list. That is a feature. A reporter can tell in one line whether a story is hers. Give her that line and you have done her first hour of work for her. Make her guess and you are gone.
The second half of positioning, the difference, has to be something you can defend when a reporter pushes back, because she will. Everyone says they are easier to use, she will tell you, so why are you actually different? If the honest answer is we are not, really, we are just newer, that is worth knowing now, in private, before you have built a campaign on a claim that collapses on the first call.
Real difference lives in something concrete: a mechanism nobody else uses, a customer nobody else serves well, a result you can show, a constraint you turned into an advantage. We are the only payroll tool that files directly with the Singapore and Hong Kong authorities in one flow is a difference. We are passionate about customer success is wallpaper. The test is whether a competitor could paste the sentence onto their own site without lying. If they could, you have not written positioning, you have written filler.
Once you have the audience and the difference, freeze them into one line and make everyone use it. The founder on a podcast, the landing page headline, the first line of every pitch, the company description a model scrapes. Same shape, same words, every time. Positioning that lives in one person's head drifts the moment a second person describes the company. Positioning on the page, repeated, is what a reporter can place and a model can retrieve. The next chapter turns that single line into the handful of claims that carry it, with the proof attached.
Takeaway: Pick one audience and one real difference, write them as a single line a stranger can repeat, and let that line decide which desk and reporter you belong to.
Chapter 3
Positioning gives you one line. Messaging gives you the few claims that carry that line into the world, each with proof attached, written down in one place everyone draws from. The tool for this is old and reliable, and PR teams have used it for decades. It is called the message house.
Picture a simple house drawn on one page. The roof is your positioning line, the single sentence from the last chapter. Underneath it stand three or four pillars, and each pillar is one claim you want a reader, a reporter, and a model to believe about you. The floor the whole thing rests on is proof: under every pillar, the specific evidence that makes the claim true.
Three pillars is the right number for most companies. Two feels thin. Five starts to blur. Each pillar should be a complete, concrete claim, not a theme. Fast is a theme. Onboards a new restaurant in a single afternoon is a claim. The difference is that a claim can be proven or disproven, and a reporter can check it.
A claim with no proof under it is a hope. The proof is what turns your messaging from marketing into something a reporter can repeat and a model can cite. For each pillar, write the evidence directly beneath it. Proof comes in a few forms, and the stronger ones sit outside your own opinion:
If a pillar has no floor, you have two honest choices. Build the proof, or drop the claim. Pitching a claim you cannot back is the fastest way to lose a reporter for good, because her credibility is on the line when she prints it, and she knows it even when you forget.
The message house only works if it lives in one document that every surface pulls from. One page. The roof, the pillars, the proof, and nothing else. Not a sixty-slide brand deck nobody opens. A single sheet the founder, the person writing the landing page, and the person sending pitches all read from the same morning.
When that document exists, consistency stops being a matter of willpower and becomes automatic. The landing page headline is the roof. The three sections below it are the pillars. The pitch leads with the roof and offers the most relevant pillar with its proof. The founder's bio uses the same words. The investor update echoes them. Nobody is inventing language on the fly, so nobody drifts.
This is where the message house earns its keep in the AI era. A model deciding whether to name you is, in effect, checking whether the same claim shows up in the same shape across the places it can see: your site, your founder's profile, a news article, a directory, a podcast transcript. Repetition of a consistent, checkable claim reads as truth to a model, the same way it does to a person. Say the same specific thing everywhere and you give the model a stable fact to attach to your name. Say five different things and you give it noise, so it quietly drops you from the answer and names a competitor who stayed on message.
This is the quiet reason scattered messaging is so expensive now. In the old world, inconsistency across your surfaces was untidy and most humans never noticed. In the new world, a model samples all of those surfaces at once and notices immediately. The message house is how you make sure every sample returns the same answer.
Start from the positioning line and ask: what are the three things a skeptical buyer must believe to choose us? Those are your pillars. Phrase each as a concrete, falsifiable claim. Then, under each, write the single strongest piece of proof you have today. Be honest about which proof is thin. A thin floor is a to-do: go get the number, get the case study, get the integration listed, before you lean on that pillar in a pitch.
Keep the language plain. A pillar that needs jargon to sound impressive is usually hiding the fact that it says nothing. If you cannot explain the claim to a smart friend outside your industry, a reporter will not be able to explain it to her readers, and a model will not find a clean way to phrase it either.
Once the house stands, protect it. Every new page, pitch, and profile starts by opening this one document. When reality changes, when you earn a new number or a bigger customer, update the house first, then let the change flow outward to every surface. The house is the source. Everything else is a copy. The next chapter is about the one thing the house does not capture on its own: how all of it should sound.
Takeaway: Put three provable claims and their proof on one page, pull every pitch and page from it, and let that consistency be the thing both reporters and models reward.
Chapter 4
Positioning is what you say. Messaging is the claims that carry it. Voice is how all of it sounds. It is the difference between two companies making the identical claim, where one sounds like a person who has done the work and the other sounds like a press release written by a committee. Buyers feel that difference in a second, and so, increasingly, do the readers who decide whether to trust you.
Founders treat voice as the last coat of paint. It is closer to the foundation. Voice is what makes your content recognizable as yours, and recognizable is what earns trust over time. When your home page, your pitch, your founder profile, and the quote you give a reporter all sound like the same human being, a reader builds one coherent picture. When they sound like four different companies, the reader builds nothing, and no single impression is strong enough to remember.
There is a practical version of this too. A reporter who interviews you will quote you. If your website is stiff corporate language and you, on the phone, are sharp and specific, the quote she prints will clash with the page she links to. The reader clicks through and feels a small dissonance. Keep your voice steady across both and the whole thing reinforces: this person sounds like their company, and their company sounds like a real thing.
Left alone, every company drifts into the same beige dialect. We leverage cutting-edge technology to deliver best-in-class solutions that empower stakeholders to achieve their goals. That sentence could belong to a bank, a shoe company, or a missile maker. It says nothing, it sounds like everyone, and a model trained on a billion such sentences cannot use it to tell you apart from anyone else.
This failure mode is easy to miss because it feels safe. Corporate voice is what you reach for when you are nervous and want to sound legitimate. It does the opposite. It makes you sound like every other nervous company. Specificity is what sounds legitimate. A founder who says we cut a restaurant's payroll run from three hours to twenty minutes sounds more credible than one who empowers operational efficiency, because the first one has clearly seen the actual work.
You do not have to invent a voice. You almost always already have one, and the work is to notice it and let it onto the page. A few ways in:
A real voice has a point of view. It is willing to be specific about who it is for and blunt about what it is against. That edge is what makes a founder interesting to a reporter, who is looking for someone with an opinion rather than a spokesperson reading talking points. It is also what makes you memorable to a reader who met you once in an article and has to recall you later, when a model asks them, in effect, who to trust.
Voice has to be consistent across surfaces, though consistent does not mean identical in register. You can be more formal in a regulatory filing and looser on social, the way a person is more formal in a meeting and looser at dinner while staying recognizably themselves. What stays constant is the core: the vocabulary, the level of directness, the point of view, the refusal to hide behind jargon. A reader should be able to cover the logo on any piece of your writing and still know it is you.
This matters for an Asia-first company in particular. A consistent voice has to survive translation. If you operate in English, Bahasa, and Mandarin, the voice is not the English phrasing, it is the directness and the point of view underneath it, which a good translation carries and a literal one loses. Write the core plainly enough that it crosses languages without turning back into beige.
Voice belongs in the same source-of-truth document as your messaging. A short voice note is enough: three adjectives for how you sound, a list of words you use, a list of words you ban, and two or three example sentences written the right way. We sound direct, specific, and warm. We say use, never utilize. We never say leverage, synergy, or solution. Here is how we describe onboarding. That half page keeps every writer, including you at 11pm before a deadline, on the same voice.
Now you have the three pieces of a clear brand: a position, the provable claims that carry it, and a voice that makes it sound like you. On their own they are a document. The next chapter puts them to work, flowing down into the pages, profiles, and pitches that reporters and models actually read.
Takeaway: Find the plain, specific way you already talk about the work, hold it steady across every surface, and let that consistent human voice be what readers and reporters recognize.
Chapter 5
A clear brand is not a document you file and admire. It is the top of a funnel, and everything below inherits from it. Positioning, messaging, and voice sit at the top. Below them sit the surfaces the world actually touches: your landing pages, your founder profile, your pitches, the coverage that results, and the answers an AI gives when someone asks about your category. Each layer takes its substance from the layer above. No layer invents a new claim. That single rule is what makes a brand compound instead of leak.
The home page is the first inheritance. The headline is the positioning line, close to word for word. The sections below are the pillars from the message house, each with its proof shown rather than asserted. The copy is in your voice. Done this way, the page is not a creative project every quarter. It is a rendering of the brand, and it changes only when the brand does.
This discipline pays off twice. A visitor gets a page that says one clear thing. And a model that reads the page gets the same clean claim it will later find echoed in your profile and your coverage. The page becomes one more consistent sample, reinforcing the others.
Founders underrate their own profile. In the AI era it is one of the most sampled surfaces you own. When a reporter checks you out before a call, she reads your bio. When a model assembles a view of your company, your founder profile is often where it anchors the human story. If that profile uses different language than your home page, you have handed both of them a contradiction.
So the profile inherits too. The bio opens with the positioning line, in the first person and in your voice. It states what you do and who for, then one line of why you, the credibility that makes the difference believable. The same claims, the same words. A founder profile that matches the company is a small thing that quietly resolves a dozen downstream doubts.
Now the pitch almost writes itself, which is the whole promise of doing the work in order. A pitch is three moves: the positioning line, so the reporter can place you in one sentence; the single pillar most relevant to her beat, with its proof, so she has a story and the evidence in the same breath; and a reason it matters now, the hook that makes it this week's story and not a timeless advertisement.
Because every piece comes from the house, the pitch agrees with the page she will click to and the profile she will check. There is nothing to reconcile. A fintech reporter at The Business Times gets the payments pillar. A deals reporter at DealStreetAsia gets the growth pillar. Same brand, same claims, aimed at the right desk. The founder who skipped the brand work cannot do this. They are writing each pitch from scratch, guessing at language, and contradicting their own site without knowing it.
When a reporter writes about you, she compresses you into a line or two for her readers. If your brand is clear, she compresses it correctly, because you handed her the line and she had no reason to invent her own. The article describes you the way your page does. That alignment is worth more than the traffic. It means the coverage reinforces the brand instead of introducing a fourth version of your company into the world.
If your brand is fuzzy, the reporter guesses, and her guess becomes a published, authoritative description that may not match anything else you have said. Now a model has two conflicting sources, one of them a credible news outlet. You have made your own consistency problem worse by getting press. That is the trap of pitching before you are clear: the coverage you wanted undermines you.
The bottom of the funnel is the answer a model gives when a buyer asks it who to trust in your category. By the time you reach this layer, you are not writing anything new. You are collecting the interest on everything above. The model has seen your page, your profile, and your coverage. If they all said the same specific, checkable thing, the model has a stable claim to repeat, and it names you with confidence. If they disagreed, it has noise, and it names someone clearer.
This is why the funnel has to be built from the top down. You cannot optimize the AI answer directly. There is no page you can edit that is the answer. The answer is a reflection of everything upstream. The only way to change it is to make the upstream consistent, which is to say, to fix the brand and let it flow.
One rule governs the whole funnel: a lower layer may select from the brand above it and shape it for the moment, and it may never invent a claim the brand does not contain. The pitch chooses which pillar to lead with. The landing page decides which proof to show first. None of them gets to make up a new promise because it sounds good in the moment. The second a layer invents, the consistency breaks, and the break shows up at the bottom as a model that will not name you. Keep the rule and the whole stack pulls in one direction. The last chapter shows how that direction becomes coverage, then citations, and how to run the three as a single loop.
Takeaway: Build the funnel from the top down, let every page, profile, and pitch inherit from the brand and invent nothing new, and the AI answer at the bottom takes care of itself.
Chapter 6
Everything so far has pointed at one outcome: a clear brand that turns into coverage, and coverage that turns into citations, running as a single loop instead of three disconnected projects. This is the part most companies get wrong by keeping brand, PR, and AI visibility in separate boxes, run by separate people, measured in separate ways. They are one system. Cites X was built to run them as one.
Earned media does something your own pages cannot. When The Business Times or e27 describes you, a third party has staked its credibility on your claim. That is why coverage carries more weight than advertising with both humans and models: it is a check you did not write yourself. The check only clears if the reporter prints the claim you actually want to own. That is the payoff of the brand work. You are not hoping she invents a flattering line. You handed her a clear one, backed by proof, and she had every reason to use it.
So the sequence from the earlier chapters holds all the way down. Clear position, provable claims, consistent voice, aimed at the right desk. The coverage that results says the same thing your site says, in a more credible venue. Now the claim exists in two places that agree, and one of them is a news outlet.
Here is where the loop closes in the AI era. When a model answers who are the serious players in cross-border payroll for Asian startups, it leans on sources it can trust and verify. A news article sits near the top of that list. If your coverage states your positioning clearly, that article becomes the checkable source the model attaches to your name. Your own site said it. A reporter confirmed it. The model now has agreement across independent sources, which is the signal it needs to name you in the answer with confidence.
This is the quiet reason brand, PR, and AI citation are one loop and not three errands. The brand produces the claim. PR gets a credible third party to verify the claim. The verified claim is what earns the citation. Pull any link out and the chain breaks. Fuzzy brand, and there is no claim to verify. No PR, and the claim never gets third-party backing. No attention to AI visibility, and you never learn whether the loop is working.
A loop is only worth the name if it feeds back. After coverage lands, you watch two things. First, the ordinary PR outcomes: who replied, what published, which pillar and which desk worked. Second, the AI-citation reading: when someone asks the assistants about your category, do you appear, how are you described, and is the description the one you built. That reading tells you whether the brand is landing where buyers now look, or whether a competitor owns the answer.
Both signals feed back to the top. If a pillar keeps earning coverage, lean into it. If reporters keep misdescribing you, your positioning line is not as clear as you thought, so fix the roof and let it flow down again. If a model names a competitor and not you, find the claim they are consistent on where you are fuzzy, and close the gap. The brand is not carved once. It is tuned by what the coverage and the citations tell you, cycle after cycle. This is the loop Cites X runs, joining brand strategy, earned media, and AI-citation visibility so each one informs the others instead of guessing in isolation.
The common alternative is a brand agency, a PR firm, and a new AI-visibility tool, bought separately. Each optimizes its own box. The brand agency delivers a deck. The PR firm chases clips. The AI tool reports a score. None of them owns the claim end to end, so the claim drifts between them, and drift is the one thing that kills you with both reporters and models. Running the loop as one system keeps a single claim intact from strategy to citation. For a lean team that is also the only affordable way: you do the brand work once, properly, and every downstream surface inherits it instead of paying three vendors to reinvent it three ways.
Do not try to build the whole loop at once. Build the roof, because everything hangs from it. This week, write your one-line positioning: who you serve, what you do differently, in a sentence a stranger can repeat. Read it to three people outside your company. If they nod and play it back, you have your line. If they hesitate, rewrite it until they stop hesitating.
Then do the cheapest, highest-leverage thing with it: make your home page headline and your founder bio say exactly that line, in your own plain words. Nothing else yet. Just those two surfaces, agreeing, in your voice. You have now started the funnel at the top and made your two most-sampled surfaces consistent. From there you add the three pillars and their proof, then the pitch, then the targeted outreach. Each step inherits from the one above, and each is easier because you did the first step first.
That is the whole argument of this guide, reduced to one action. Brand before PR. Get clear, in one line, before you spend a single pitch. The coverage gets easier, the citations follow, and the company that was unplaceable becomes the one the reporter can place and the model is glad to name.
Takeaway: Run brand, PR, and AI citation as one loop that learns from every placement, and start this week by writing the one line everything else will inherit.
Cites X runs the whole loop: build the brand, win the coverage, and become the answer AI gives.
Zugang anfragen → More Field Guides →