Backed by Y CombinatorThree documents and three sheets that measure an executive's real voice from their own transcripts, so a first draft already sounds like them.
Free download · No account needed
What you get
Pack of 6 files: 3 documents, 3 spreadsheets
Executive Thought Leadership Content PlanVoice Profile
Five dimensions, each counted from the executive's own transcripts and memos rather than described in adjectives.
| Average Sentence Length | |
|---|---|
| Contractions per 1,000 Words | |
| Direct-address Rate | |
| Absolute-certainty Rate | |
| Ownership-language Rate |
A voice guide written as adjectives cannot be checked against. A draft either falls inside these five measured bands or it does not.
Most executive thought-leadership plans start with a voice guide built from adjectives: a few words like "confident" and "approachable," plus a short list of dos and don'ts. Adjectives cannot be checked against, so two reviewers approve two different drafts and the executive still rejects both. It matters because 64% of target buyers and 63% of hidden buyers spend more than an hour a week reading or watching thought leadership before a deal reaches a rep. A piece that does not sound like its author gets noticed for that before anyone reads what it argues.
Naomi Petrakis, the fictional CEO of an industrial-sensor company called Verrallis, gives this pack its worked example. Her Voice Profile is measured from 20,036 words of her own talks and internal memos: 13.6 words per sentence, 30.3 contractions per 1,000 words, and zero instances of "leverage," "unlock," or "robust." Four early drafts, written against a generic "confident, approachable" brief before this profile existed, ran to 24.8 words per sentence with almost no contractions and five times her real rate of unqualified certainty. Naomi's own markup changed 33.8% of those drafted words.
The next four drafts, held to those same five numbers instead of a paragraph of adjectives, came back only 10.2% changed. The voice-related share of editing fell from 26.9% to 3.4% of drafted words, while the factual-edit share held flat near 6.8%. Topic Bank then ranks her unpublished opinions by Engagement by Topic's resonance rate, shares per 1,000 views, which is why category critique, published more than any other tag, ranks last of six. The same sourced, dated opinions can also supply an authentic quote for a press release instead of a generic one.
Voice Profile20,036 words of Naomi's own talks and memos, measured against four first-round drafts written from a generic brief.
| Dimension | Her corpus | First-round drafts |
|---|---|---|
| Sentence length | 13.6 words | 24.8 words |
| Contractions / 1,000 words | 30.3 | 2.5 |
| Direct address / 1,000 words | 14.4 | 5.9 |
| Absolute-certainty / 1,000 words | 0.3 | 5.4 |
| Ownership language / 1,000 words | 3.4 | 1.1 |
| Batch | Executive edit rate | Voice-related share | Factual share |
|---|---|---|---|
| 4 drafts, generic brief | 33.8% | 26.9% | 6.9% |
| 4 drafts, held to the profile | 10.2% | 3.4% | 6.7% |
The factual-edit share barely moved, 6.9% to 6.7%, because nothing about a voice profile makes a fact more accurate. That flat line is the evidence the 87% drop in voice-related editing is real.
Engagement by Topic11 published pieces across 18 months, grouped by tag and ranked by shares per 1,000 views rather than raw views.
| Tag | Pieces | Views | Shares | Shares / 1,000 views |
|---|---|---|---|---|
| Candid failure | 2 | 6,750 | 132 | 19.56 |
| Business model philosophy | 2 | 3,950 | 76 | 19.24 |
| People and culture | 1 | 2,900 | 51 | 17.59 |
| Data and customer trust | 1 | 610 | 8 | 13.11 |
| ROI and measurement | 2 | 5,140 | 60 | 11.67 |
| Category critique | 3 | 13,280 | 60 | 4.52 |
| All 11 pieces | 11 | 32,630 | 387 | 11.86 |
Category critique has the most published pieces and the lowest rate, by a factor of four against the top tag. It is the easiest topic to write and the one the audience responds to least.
Publishing CalendarSix slots on a three-week cadence, sequenced from Topic Bank's highest-priority open entries.
| Topic | Format | Review due | Publish | Status |
|---|---|---|---|---|
| The maintenance schedule nobody changed | Mar 9 | Mar 16 | Published | |
| We turned down a per-sensor deal | Mar 30 | Apr 6 | Published | |
| Why I stopped asking about architecture | Byline | Apr 20 | Apr 27 | Published |
| The pilot that worked, killed anyway | May 11 | May 18 | Published | |
| Three numbers before a board meeting | Podcast | Jun 1 | Jun 8 | In Review |
| What happens to your data on exit | Jun 22 | Jun 29 | Queued |
Draft-review reminders fire ahead of each Review Due date, sized to how long this executive actually takes to turn markup around.
Five measured dimensions, sentence length, contractions, direct address, absolute-certainty rate, ownership language, counted from real transcripts and memos, never guessed at.
Only positions the executive has actually stated, each sourced and dated, so a draft never rests on an opinion invented to fill a topic slot. The same register can back an award or speaking submission's strongest claim just as well as a byline, or supply the new evidence an analyst briefing needs once a published note has already gotten the story wrong.
Two finished pieces written against the measured bands, so the next draft has a real example to match rather than a description to interpret.
Every open point-of-view entry ranked by its tag's real engagement history, so the easy topic to write does not outrank the one that resonates.
Past published pieces grouped by tag and ranked on shares per 1,000 views, the active-response rate, rather than on passive views alone.
A realistic cadence with draft and review dates sized to how long this executive actually takes to turn markup around, plus review reminders.
Hand over talk recordings, podcast appearances, and internal memos the executive wrote themselves, or start with a short recorded interview if little exists yet.
River counts sentence length, contractions, direct address, and certainty language across the real corpus, and flags any word that never appears at all.
Point-of-view Register only accepts positions the executive has actually voiced. A topic with no register entry becomes an interview question instead.
When a draft comes back marked up, River splits the changes into voice-related and factual, which is what proves whether the profile is still working.
Yes, with no signup and no card, and the download is the whole pack rather than a preview of it. Edit with AI is the other path: send River the executive's own material and it measures their real profile instead of this worked one. More packs sit in the template library.
Three Word documents and three CSV sheets in one zip. The documents open in Word, Pages, and Google Docs, and the sheets open as real spreadsheets in Excel, Numbers, or Google Sheets rather than a table flattened onto a page.
The usual calendar names topics and dates and tells you to write in an authentic voice. This one measures that voice first, as five numbers from real transcripts, and ranks topics by what the executive's own audience has already shared rather than by what sounds timely.
Talks and internal memos count as much as published writing, and most executives have more of both than they think. Naomi's worked profile leans on five conference or podcast transcripts plus two dozen internal memos, not on a large existing body of bylined articles.
Judgment is exactly what a corpus-measured profile protects, because judgment applied draft over draft compounds small, individually reasonable edits until the voice has drifted. Checking each new draft against the original transcripts, not the last approved one, is what catches that before the executive does.
Yes. A pitch angle is only as credible as the opinion behind it. The media list and pitch pack works better fed by a register of positions the executive has actually stated than by a generic claim written to fit the news.
It installs this pack as a private workspace and asks for talk transcripts, memos, or recordings. It measures the five dimensions, mines real stated positions into the register, and drafts the next piece against those bands rather than an adjective brief.
Written by
Co-founder & CEO, River
Chandler has spent the last six years working in AI, building tools for millions of users and Fortune 100 companies. In 2025, he co-founded River to build a new kind of operating system for AI agents.
Send talk transcripts, memos, or recordings. River counts the five dimensions, mines real stated positions, and ranks what to draft next by what has already resonated.
Edit with AI