Operating principles

Five rules. Everything else is just product.

We don't build magic. We build infrastructure. The principles below aren't marketing — they're the rules we design inside, and the ones we refuse to break when a shortcut would be cheaper.

—The five

Humans remain authors of record.

Every answer cites the person who wrote it. A Representative is a delivery mechanism, not an author. No wrap, no voice — the system won't speak on someone's behalf without something they've signed.

Why it exists

Accountability cannot be delegated to software. When a decision is communicated, a human has to own it.

What it prevents

"The AI said so" excuses, quiet loss of ownership, and the slow drift where people stop writing because the tool fills in for them.

Written, not observed.

StandIn only knows what was explicitly published. It doesn't scrape activity, read DMs, infer intent from Slack patterns, or measure who types when.

Why it exists

Surveillance tools kill the trust that makes async work possible. A system optimized to watch people will be rejected by the people it watches.

What it prevents

Monitoring creep, manager dashboards that track presence instead of progress, and the quiet coercion that comes with being observable by default.

Silence is a valid answer.

When no one has written the answer, the Representative says so — and points to the owner if there is one. It doesn't fill the quiet with a plausible guess.

Why it exists

A tool that's right 90% of the time and confidently wrong 10% of the time is unusable for work that matters. Silence protects the value of every answer that isn't silence.

What it prevents

Hallucinations dressed as status, decisions made on inferred context, and the steady erosion of trust that follows the first confident bluff.

Refusal builds trust.

Some queries get declined by design — surveillance, sentiment, DMs, inference about people. Refusal isn't a bug or a scope limit. It's the architecture.

Why it exists

Some answers shouldn't exist even when the data technically could produce them. The refusals are what make StandIn safe to put in front of a whole team.

What it prevents

Scope creep into HR monitoring, prompt-engineering around ethical guardrails, and pressure from well-meaning stakeholders to "just answer this one."

Continuity over presence.

The goal isn't that more people are online more hours. It's that the work keeps moving when they aren't. We optimize for the handoff, not for attendance.

Why it exists

Distributed teams break when "always online" becomes the workaround for missing context. Continuity of record is the actual fix.

What it prevents

After-hours pings, standing meetings that exist to repeat what was already written, and a culture where availability is mistaken for contribution.

If these rules sound right, so will the product.

See how it works Back to home

The full protocol →