Voice Configuration
Voice configuration controls how Pilot writes — tone, formality, confidence, citation density, and brevity. The same knowledge base can produce different content for different audiences by changing the voice settings.
Your Voice, Not Pilot's
Pilot doesn't have a default writing style. It has yours. Voice configuration is the set of parameters that control how Pilot writes — not what it says (that comes from your knowledge base), but how it says it.
This matters because the same knowledge needs to sound different depending on who's reading it. A client newsletter doesn't read like a research brief. A LinkedIn post doesn't read like a regulatory filing. Same facts, different voices. Voice configuration is how you make that distinction.
The Parameters
Voice settings are controlled through sliders in the Pilot console. Each one shapes a specific dimension of the writing.
Brevity
How much room does Pilot take to make a point? At one end: tight, compressed prose that covers ground quickly. At the other: expansive writing that develops ideas with examples and context. A social media channel needs compression. A deep-dive article needs space.
Formality
How does Pilot address the reader? Casual voice uses contractions, direct address, shorter sentences. Formal voice uses complete constructions, measured phrasing, academic precision. A trade publication aimed at practitioners lives in the middle. A policy brief for legislators leans formal.
Confidence
Does Pilot hedge or assert? Cautious voice qualifies claims: "research suggests," "evidence indicates," "one interpretation is." Assertive voice states positions directly: "the data shows," "this approach works," "the standard requires." Both are valid — the right setting depends on your domain and audience expectations.
Citation Density
How heavily does Pilot source its claims? Light citation gives the reader a clean reading experience with occasional references. Heavy citation attributes nearly every substantive claim to a specific source document. A newsletter aimed at general readers benefits from lighter sourcing. A technical report for subject-matter experts benefits from dense attribution.
Tone
The emotional register of the writing. Warm voice feels approachable and conversational. Neutral voice is straightforward and professional. Authoritative voice carries institutional weight. There are additional dimensions — you have eleven parameters total — covering aspects like directness, technical depth, and reading level.
Same Knowledge, Different Voice
Here's what voice configuration looks like in practice. The same information about thermal bridging in residential construction, drawn from the same source documents, written in two different voices:
Voice A — Advisory, moderate formality, assertive confidence:
Thermal bridging at stud locations accounts for 20-30% of a wall assembly's total heat loss. Standard fiberglass batt insulation between 2x6 studs leaves the studs themselves as direct conductors. Continuous exterior insulation — even one inch of rigid foam — breaks those thermal bridges and measurably improves whole-wall R-value.
Voice B — Editorial, casual formality, measured confidence:
The studs in your walls are working against you. While insulation fills the cavities between framing members, the wood itself conducts heat at a rate that can undermine 20-30% of your wall's insulating value. Builders call this thermal bridging, and it's one of those details that doesn't show up in a spec sheet but shows up in every energy bill. Adding continuous insulation on the outside of the sheathing is the most direct fix.
Same facts. Same source documents. Same citations available. Different reading experiences for different audiences. The advisory voice serves a contractor who needs the technical specifics. The editorial voice serves a homeowner who needs to understand why it matters.
Multiple Voices Per Tenant
A single Pilot tenant can maintain multiple voice configurations. This is how organizations serve different audiences from the same knowledge base.
A trade publisher might configure three voices: an editorial voice for feature articles, a technical voice for reference guides, and a brief voice for newsletter items. Each voice is tied to a specific channel, so articles generated for the newsletter automatically use the newsletter voice.
A consulting firm might configure a formal voice for client-facing reports and a more conversational voice for thought leadership blog posts. Same expertise, different registers, different channels.
Voice configurations are independent of each other. Changing the newsletter voice doesn't affect the report voice. Each one evolves as you tune it — adjusting a slider, reviewing the output, adjusting again until the writing matches what you'd produce yourself.
Why Voice Matters
Most AI writing tools have a "tone" setting or a style prompt. That's a sentence or two of instruction before each generation. Pilot's voice configuration is more than a prompt — it's a persistent editorial framework that shapes every article generated through that voice.
The difference shows up in consistency. A one-off prompt produces text that varies with each run. A voice configuration produces text that holds a steady editorial posture across dozens or hundreds of articles. Your tenth article sounds like it was written by the same editorial team as your first — because the same voice parameters shaped both.
This consistency is what lets organizations put their name on Pilot's output. The content doesn't just contain your knowledge — it sounds like your organization. Readers who know your work recognize the voice. That's not a nice-to-have for publishers and knowledge businesses — it's a requirement.
Configuring Your Voice
Voice settings live in the Pilot console under Settings. You adjust sliders, see a preview of how the current configuration would shape sample content, and save. The console also shows a drift score — a measure of how much your voice settings have changed since your last published content, so you know when a reconfiguration might produce noticeably different output.
For a step-by-step walkthrough of setting up your first voice, see Getting Started. For a broader view of how voice fits into Pilot's editorial system, see How Pilot Works.
Last updated March 3, 2026