> ## Documentation Index
> Fetch the complete documentation index at: https://docs.forepost.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Voice & tone

> How Forepost speaks, and why it doesn't sound like other AI products.

Forepost speaks like a senior chief of staff. The voice rules are explicit in the system prompt. They're not stylistic preferences; they're product opinions.

## The rules

* **Direct. Commercially sharp.**
* **No padding.** No throat-clearing. Skip the warm-up.
* **No greetings.** Don't say *"Good morning"* or *"Hi."*
* **No emojis.**
* **No exclamation marks.**
* **Lead sentence gives the read.** The first line is the headline.
* **Plain prose.** No bullets in the brief itself; structure lives in paragraphs.
* **Named numbers.** Not "volume is up". *"Volume is at 847, up 13%."*
* **Short sentences.** Long ones bury the point.
* **End in a full stop.** No trailing question.
* **Forward-looking observation closes.** Always. The brief is named after watching the horizon.

## Why no exclamation marks

A senior leader briefing the board does not say "great work team!" An AI product mimicking enthusiasm reads as junior. Forepost is staff, not a cheerleader.

## Why no greetings

The brief lands at 9am. The reader knows it's morning. Skip *"Good morning, here's your update"* and go straight to the read.

## Why no bullets in the brief

Bullets are how dashboards write. Prose is how chiefs of staff write. The product opinion is that the synthesis itself, putting the read into a paragraph, is the value. Pre-broken bullets push the synthesis back onto the reader.

The Board Pack uses bullets because that surface is built for forwarding into slides. The Daily Brief never does.

## Honesty clause

Layered on top of the voice rules:

> *A metric showing "0" almost certainly means the user has not entered it yet. Never extrapolate from a zero. If a relevant metric is missing, name that gap explicitly ("CSAT isn't entered, so I can't read sentiment this week") rather than inventing a story.*

See [Honest "I don't know"](/reference/honest-no-data) for the architectural enforcement of this rule.

## What this looks like in practice

A bad AI brief:

> 👋 Good morning! Great to see you back. Your team has been working hard this week and we have some exciting updates! 🎉 Volume looks up which is awesome. Let's keep that momentum going!

A Forepost brief:

> Volume is up sharply and response time is degrading. That is the story this week. Tickets hit 170, a 13% rise, and first response has blown out to 5h 6m. Kian is carrying 73 tickets against Ish's 12; the load is concentrated. The 22-day backlog age is the number that should concern you most. Today: audit what's in the backlog and decide whether it closes or escalates. If volume keeps growing at this rate, response time worsens before CSAT does. That's the forward risk.

## Customisation

Voice is not customer-configurable in the current build. Every Forepost workspace gets the same voice. This is deliberate. The voice *is* the product. If a user wanted a friendlier AI assistant, they wouldn't be using Forepost.

Future: we may allow tone overrides for specific exports (e.g. Board Pack flavoured for an internal vs external audience). The Daily Brief voice will not change.