▎ OpenRift — Operating Layer — 01
Slack-first.Agent-driven.Audit-trailed.
human-in-the-loop · every external action
OpenRift · operating layer · Singapore
T+00:12watched inbox.unread → classified quarterly-statement
T+00:14drafted reply.v1 → posted #advisory-ops · awaiting approval
T+00:17awaiting human review
T+00:23approved by partner@firm · sent
T+00:23archived thread.client-doc-q1-2026 · evidence link logged
WHAT IT IS
An operating layer for the firms in between.
Between off-the-shelf AI tooling and bespoke enterprise builds.
OpenRift is an AI operating layer for the firms that fall between off-the-shelf AI tooling and bespoke enterprise builds. It reads the tools the firm already runs, sits in the channels the team already uses, and produces drafts that a partner approves before anything ships.
Who it's for: boutique professional-services firms — five to fifty staff, with regulated or document-heavy operating workflows, who have outgrown spreadsheets and inboxes but cannot justify a six-figure enterprise build.
WHAT'S AUDIT-TRAILED
Every external action runs human-in-the-loop.
The agent prepares; a partner approves; the system ships.
Every external action runs human-in-the-loop. The agent prepares; a partner approves; the system ships. Drafts land in a draft folder, never in a client's inbox, unless a partner clicks approve. The approval is logged.
The compliance posture is not bolted on — it is the architecture.
Operating from Singapore.