A full-pipeline playbook for operators who want media coverage on repeat
Earned media is still the highest-trust asset in marketing, and Qwoted has become one of the fastest routes to it: journalists post source requests, experts pitch, and the best answer wins. The problem is that “best” is decided in hours, not days — and most operators lose on speed, relevance, or both.
This guide shows you how to run Qwoted as a system, with Claude as the leverage layer at every stage: qualifying requests, drafting pitches in your expert’s authentic voice, following up, and turning every win into more wins. It’s written for high-intent operators — founders doing their own PR, in-house comms leads, and agency teams running Qwoted for multiple clients.
One principle governs everything below: Claude drafts, the expert owns. Every quote that goes out must be something your expert actually believes and can defend on a call. AI-flavored filler is the fastest way to burn a journalist relationship. Used correctly, Claude doesn’t make your pitches sound like AI — it makes them arrive faster, tighter, and more on-point than anyone pitching manually.
How Qwoted Works (60-Second Refresher) #
Journalists post source requests(“Calls for Experts”) describing the story and the source they need.** You unlock a requestto see the journalist’s identity and pitch — unlocking spends a pitch credit whether or not you submit. Your pitch lands in the journalist’s Qwoted inboxinstantly, with a real-time email alert on their end. Plans:**the free tier includes 2 pitches/month with a 2-hour delay before you can pitch; Pro (~$99/month) includes 35 pitch credits, no delay, pitch intelligence, and profile-view data. Custom team plans add reporting and account management.Two things Qwoted’s own data and power users agree on: pitches of roughly 7–12 sentences get the best response rates, and speed matters — get your pitch in well within 12 hours of the request posting, even if the deadline is days away.
The economics follow directly: every credit is a paid shot, the delay on the free tier is a real handicap in a speed game, and the winning pitch is short, specific, and early. That’s exactly the game Claude helps you win.
Stage 0: Build Your Claude Foundation (One-Time Setup, ~1 Hour) #
Before you respond to a single request, give Claude the context it needs to write like your expert instead of like a chatbot. Do this once per expert or client; it pays back on every pitch afterward.
Create a dedicated Claude Project (or a persistent context doc if you’re working in another surface) containing:
Expert dossier. Name, title, company, one-paragraph bio, credentials, notable numbers (“scaled X to $Y,” “analyzed Z transactions”), past media placements, and the 5–8 topics they can credibly speak on. Be honest about edges — Claude should know what your expertcan’tspeak to, so it flags bad-fit requests instead of stretching.Voice sample bank. Paste 3–5 examples of the expert actually talking: podcast transcript excerpts, past published quotes, internal memos, even Slack messages. This is the single highest-leverage input. Claude mimics voice from examples far better than from adjectives like “conversational but authoritative.”Opinion inventory. The expert’s 10–15 contrarian or distinctive takes, stated bluntly. (“Most founders raise too early.” “The 60/40 portfolio isn’t dead, it’s misunderstood.”) Journalists don’t need five good points — they need one sharp one. This inventory is where sharp points come from.Pitch rules. Your non-negotiables, written as instructions: 7–12 sentences, journalist’s first name in the greeting, expert credibility established in the first two lines, one insight per pitch, a quotable line the journalist can lift verbatim, availability and contact at the end, no em-dash-heavy AI cadence, no “I hope this finds you well,” no “in today’s fast-paced world.”Do-not-say list. Compliance constraints, competitors not to name, claims legal won’t clear, topics that are off-limits. For regulated industries (finance, health, legal), this section is mandatory, not optional.
Agency variant: one Project per client. Never let voices cross-contaminate. A pitch drafted in the wrong client’s voice is worse than a slow pitch.
Stage 1: Monitor and Triage (Daily, 10 Minutes) #
Qwoted’s matching surfaces relevant requests via real-time alerts and a daily opportunities email. Your job is to compress “scan everything” into a ruthless daily triage.
The workflow: paste the day’s batch of candidate requests into Claude with a standing triage prompt:
“Here are today’s Qwoted source requests. Against [Expert]’s dossier, score each 1–10 on: (a) genuine expertise match, (b) outlet value for our goals, (c) competition level — how many other experts can credibly answer this, (d) our angle strength — do we have a distinctive take from the opinion inventory, or would we be generic? Recommend PITCH / SKIP / STRETCH for each, with one sentence of reasoning. Flag anything with a same-day deadline first.”
Why scoring beats gut feel: every unlock spends a credit. On Pro’s 35 credits/month, you’re making ~1–2 paid decisions a day. The discipline that separates operators from dabblers is skipping the 7/10s to stay loaded for the 9/10s. A request where you’d be the fifteenth generic voice is a wasted credit even if it’s “on topic.”
The competition heuristic Claude should apply: broad requests (“looking for small business owners to comment on inflation”) attract floods of pitches — your edge must be an unusually sharp angle or you skip. Narrow requests (“need a fintech compliance officer who has implemented FedNow”) get few qualified pitches — if you match, that’s a near-automatic PITCH even at a mediocre outlet.
Deadline math: treat every request as expiring in 12 hours regardless of stated deadline. Journalists read pitches as they arrive and often mentally commit early. Triage in the morning, pitch by midday.
Stage 2: Qualify Before You Unlock (2 Minutes Per Request) #
Before spending the credit, run one more Claude pass on your PITCH-rated requests:
“Before I unlock this, pressure-test it: What is the journalist
actuallywriting about, reading between the lines of the request? What source do they probably wish existed? Does [Expert] resemble that source? What’s the one insight from our inventory that fits — and is it differentiated enough to win against likely competition? If we don’t have a distinctive angle, say so plainly.”
This is your kill switch. If Claude can’t articulate a distinctive angle in this pass, you won’t magically find one after unlocking. Skip and keep the credit.
If it survives, unlock. Now you can see the journalist’s identity — which unlocks the personalization stage.
## Stage 3: Research the Journalist (3 Minutes)
Personalization is not “Hi [FirstName].” It’s demonstrating in one line that you know what this journalist covers and why your expert fits their beat specifically.
With the journalist’s name revealed, have Claude (with web search) do a fast recon:
“Research [journalist name] at [outlet]. What’s their beat? Summarize their 3 most recent pieces. What angles do they favor — data-driven, contrarian, human-interest? Have they covered [topic] before, and what did they say? Give me one natural, non-sycophantic way to reference their work in a pitch.”
Two rules: the reference must be load-bearing (it should connect their prior work to your angle, not just flatter), and if Claude can’t find anything solid, skip the reference entirely — a forced or wrong one (“loved your piece on X” when they didn’t write X) is disqualifying. Verify any cited article actually exists before it goes in the pitch.
Stage 4: Draft the Pitch (5 Minutes) #
Now the core move. Feed Claude the full request text, the journalist recon, and invoke the Project context:
“Draft a Qwoted pitch responding to the request below. Structure: (1) Greeting with first name. (2) One–two lines establishing exactly why [Expert] is credible for
thisstory — specific numbers and credentials, not adjectives. (3) The core insight: [chosen angle from opinion inventory], developed with one supporting specific. (4) A ready-to-use quote of 2–3 sentences in [Expert]’s voice that the journalist could paste into the article verbatim. (5) Offer of availability (interview windows, timezone) and anything extra we can provide — data, screenshots, intros. Total length 7–12 sentences. Match the voice samples. Then critique your own draft as a skeptical journalist skimming 40 pitches: would you stop on this one? Revise once based on the critique.”
That self-critique step is not decoration — it reliably catches generic openers, buried leads, and hedge-words. You can also run the inverse: “Write the pitch the other fourteen experts are probably sending, then make ours the one that doesn’t sound like them.”
What makes the quote section win: journalists on deadline want their job half-done. A pitch that includes a polished, specific, immediately usable quote — with a number, a vivid comparison, or a contrarian frame — beats a pitch that merely offers to talk. This is where the opinion inventory earns its keep.
The human pass (non-negotiable): the expert (or you, with the expert’s sign-off standing) reads the draft and asks three questions. Is every factual claim true? Would I say this sentence out loud? Can I defend this position in a 20-minute interview? Edit until yes. This usually takes two minutes and is the difference between a quote and a liability. Never let a number, statistic, or client claim ship without verification — Claude drafts persuasively whether or not a claim is accurate, so verification is your job, not the model’s.
Then submit — same day, ideally same morning.
Stage 5: Follow-Up and Interview Prep #
If no reply in 3–4 business days and the deadline hasn’t passed, one short follow-up is acceptable — two lines maximum, adding something new (a fresher stat, an additional angle), never “just bumping this.” Have Claude draft it in 30 seconds. One follow-up. Never two.
If the journalist replies wanting an interview, flip Claude into prep mode:
“The journalist responded to our pitch [paste thread]. Based on their beat and recent articles, generate the 10 questions they’re most likely to ask, including the 2–3 uncomfortable ones. Draft talking-point bullets for each in [Expert]’s voice — one headline sentence plus one supporting specific per question. Flag any question that touches the do-not-say list and suggest a bridge.”
If they ask for written answers instead (common on Qwoted), draft responses with the same voice-matched process as the pitch, run the same human verification pass, and return them fast — written-answer requests are usually filling holes in a nearly finished article, so a 2-hour turnaround beats a perfect 2-day one.
Stage 6: Track, Learn, and Compound #
Coverage that isn’t tracked and reused is coverage half-wasted. Close the loop:
Track everything in a simple log (a spreadsheet Claude can maintain): date, request topic, outlet, journalist, angle used, credit spent, outcome (no reply / reply / interview / placed), and link when published. Qwoted Pro’s pitch intelligence and profile-view data feed this too.
Run a monthly retro with Claude: paste the log and ask what’s working —
“Here’s this month’s pitch log. What patterns separate the pitches that got replies from those that didn’t — topic, angle type, outlet tier, day/time submitted, pitch length? Which of [Expert]’s opinion-inventory angles are converting? What should we stop pitching, and what should we double down on? Recommend next month’s credit allocation.”
Over a quarter, this turns Qwoted from a lottery into a model: you learn which angles convert, which outlets reply, and which request types are credit traps. Feed the winning pitches back into the Project as new voice samples — the system literally improves with every placement.
Repurpose every placement. Each published quote becomes: a LinkedIn post from the expert (Claude drafts three variants), an “As featured in” addition to the website and email signature, ammunition for the next pitch’s credibility line (“recently quoted in [Outlet] on this exact topic”), and — for agencies — a client-report line item with the outlet’s reach attached.
Warm the journalist relationship. A placement is the start, not the end. Two touches Claude can draft: a short genuine thank-you after publication (with a specific reaction to the piece, not generic praise), and a note a month later flagging a relevant development on their beat with no ask attached. Journalists keep lists of reliable sources; your goal is to be on them so the next request comes to you directly — off-platform, no credit required.
The Operating Cadence at a Glance #
| Rhythm | Action | Time |
|---|---|---|
| Daily (AM) | Triage new requests with scoring prompt; qualify and unlock winners | 10–15 min |
| Per pitch | Journalist recon → draft → self-critique → human pass → submit | 10–12 min |
| As needed | One follow-up at day 3–4; interview/written-answer prep on reply | 5–30 min |
| Weekly | Update pitch log; repurpose any placements | 15 min |
| Monthly | Claude retro on the log; refresh opinion inventory and voice samples; reallocate credits | 30 min |
Total steady-state cost: roughly 90 minutes a week per expert for a system that most teams staff with a fractional PR retainer.
Failure Modes (and the Fix for Each) #
The AI-obvious pitch. Symptoms: “I hope this finds you well,” “in today’s rapidly evolving landscape,” symmetrical three-part sentences, no specifics. Fix: voice samples in the Project, the self-critique pass, and a human read. If a pitch could have been written about any expert, it wasn’t written about yours.The stretched match. Claude, eager to help, finds an angle connecting your expert to anything. Fix: the dossier’s honest “can’t speak to” list, and treating STRETCH ratings as skips by default. One off-expertise quote that goes sideways in an interview costs more than ten skipped requests.Unverified specifics. A hallucinated statistic in a published quote is a credibility fire. Fix: hard rule — every number and claim gets human verification before submit, every time.Credit bleed. Unlocking on curiosity, pitching 6/10 matches. Fix: the Stage 2 kill switch, and monthly retro data on which request types actually convert.Set-and-forget decay. The opinion inventory goes stale, pitches start repeating. Fix: the monthly refresh — new takes, new numbers, new voice samples from recent interviews.
QWOTED & Claude #
Qwoted rewards exactly three things: speed, specificity, and a source who sounds like a person worth quoting. Claude, set up properly, is a force multiplier on all three — it compresses triage from an hour to ten minutes, gets a voice-true draft in front of you in five, and turns your pitch history into an improving model of what wins. The operators who compound here aren’t the ones sending the most pitches; they’re the ones whose system makes every pitch land earlier and sharper than the field’s, and who convert each placement into the next one.
Set up the foundation this week. Run the cadence for a month. Read the retro. Then scale what converts.