Skip to content

Answer fast, in the candidate's own words.(note 1)

november listens to the race, drafts responses from what the candidate has actually said, and puts every public word in front of a person before it goes out. Campaigns that trust each other share what they make, each through its own approvals.

1Every public claim november drafts cites the speech or meeting it came from. When the record is silent, so is november.

A rebuttal being drafted in the console. One sentence's citation is open: the town hall transcript it came from, with the candidate's exact words highlighted.
A drafted rebuttal with one citation open: the sentence, and the transcript where the candidate said it. Fictional demo world: candidate Maya Okafor, Riverbend-14.

The hour after an attack decides the story

Most campaigns lose it the same three ways.

Answers come late and off-message

By the time a reply is drafted, checked and signed off, the attack has had the day. The version that finally goes out often says something the candidate never has.

Volunteers improvise

Door-knockers get the hard question on the porch and answer from memory. Supporters post whatever they could find. Nobody gave them the campaign's words.

Nobody can show where a claim came from

When a reporter asks for the source of a line, someone searches old speeches by hand. Sometimes there is no source at all.

How it works: one loop, from the record to the doorstep and back

Eight steps. Agents do the reading, drafting and checking in between; people decide at every step that matters.

  1. Ground

    Load speeches, town halls, interviews and meetings. november keeps the candidate's claims, stories and positions, each pinned to the moment she said it.

    The Truth screen: a list of the candidate's sources, a town hall, a stump speech, two strategy meetings and a radio interview, each with its date and place.
  2. Listen

    Signals from the candidate, the audience and opponents arrive clustered and ranked, so the attack that matters sits at the top.

    The Listen screen: clusters of signals ranked by priority, the top one an opponent's claim about the childcare plan, beside a live feed of posts.
  3. Respond

    An agent drafts a reply in the candidate's voice in minutes. Every factual sentence cites her own words, and guardrails check it before anyone reads it.

    The Compose screen: a rebuttal drafted sentence by sentence, each fact with its source attached, and a column of guardrails that all pass.
  4. Control

    A person reviews the draft against the campaign's KPIs, each score with its reasons. Release takes a second approver, and every override is recorded.

    The review panel: Approve and Return buttons over KPI scores for audience fit, brand and message, each with the sentences that earned it.
  5. Publish

    Approved content goes out as a versioned page with a QR code: an explainer a door-knocker can hand over, which answers questions only from the record.

    A published explainer page titled “Maya's childcare plan: the facts”, with numbered marks after each statement that open its source.
  6. Distribute

    Volunteers get a kit for each platform: approved text, ready to copy and share in one tap.

    The Distribute screen: approved Instagram, SMS, TikTok and X variants of a rebuttal assigned to three volunteers, with a board of kits sent.
  7. Measure

    Reach and engagement per piece and per volunteer, each number marked with where it came from.

    The Measure screen: totals for questions, views, clicks, shares and replies, each tagged with its source, manual or simulated.
  8. Federate

    Share a finished piece with a campaign you trust. It arrives as a frozen copy with its sources and goes through their approvals, under their brand.

    The share dialog: a rebuttal offered to a neighbouring campaign, Harbor-9, with a note to the receiver.
The loop in the demo console, one screen per step. A shared piece lands in the next campaign's record with its sources, and the loop starts again there. Fictional demo world: candidate Maya Okafor, Riverbend-14.

Each step in detail

Who it is for

Everyone who speaks for the campaign, from the candidate to the volunteer at the door.

Candidates
Nothing goes out in your name that you did not say, and every line shows where you said it. Your words, not a paraphrase.
Campaign managers
One place to see what came in, what is being answered, who approved it and what it reached. The record of every decision is already kept.
Communications directors
A cited draft in the candidate's voice, checked against your message and brand before it reaches you. You approve it, return it or rewrite it.
Field and volunteer leads
Volunteers get approved posts and explainers they can share in one tap, and a page to show at the door that says “I don't know” rather than guess.
A volunteer's phone: an approved Instagram post about the childcare plan, marked approved and not editable, ready to copy.
What a volunteer sees on a phone: approved text, ready to share. Fictional demo world: candidate Maya Okafor, Riverbend-14.

People decide. The record speaks.

november drafts quickly so people can decide carefully.

Humans decide
Agents read, draft and score. Only a person approves, releases or shares.
Grounded or silent
A factual sentence with no source in the candidate's record does not go out. Asked something the record does not cover, an explainer says so and passes the question to the campaign.
Four eyes
By default, release needs a second person. Overrides are allowed, named and kept.
Every step recorded
Who drafted, who scored, who approved and what changed, kept in an audit trail you can show.

How november keeps a campaign safe

A public explainer's question box. Asked about gun control, it answers “I don't know” and passes the question on; asked about taxes, it answers with the candidate's quoted sources.
A public explainer answering two questions: one the record covers, with sources, and one it does not, with “I don't know”. Fictional demo world: candidate Maya Okafor, Riverbend-14.

Campaigns that trust each other, sharing

Each campaign stays sovereign. When one shares a grounded piece, the other adopts it through its own approvals, under its own brand. The network sees where a piece came from and how it scored, never its content.

How the network works

The network's view of a shared rebuttal: accepted, from one campaign to another, with its scores and the note that content stays with the campaigns in the share.
The Forward Network's view of a shared rebuttal: who shared it with whom, and its scores. The text stays with the campaigns. Fictional demo world: candidate Maya Okafor, Riverbend-14.

See it work on a campaign

The demo is a working console for a fictional race. Sign in and follow one response from the candidate's record to a published, cited answer.