FAQ
Tier-2 questions, answered.
The questions sales engineers usually field on the second call. Asked and answered up front, so you don't have to chase a Calendly link to find out.
Where does our data live, and what does Omnia ingest?
Omnia is hosted on AWS in your region of choice (US-East, US-West, or EU-Central). We ingest read-only feeds from Slack, Jira, Zoom, Google Workspace, and your HRIS via OAuth. We never write back unless you explicitly approve a workflow. All ingest is encrypted in transit (TLS 1.3) and at rest (AES-256).
Which LLM do you use? Does our data train it?
Omnia runs on a SOC 2 Type II–covered LLM deployment with a zero-retention agreement: prompts and completions are not used for training, and are deleted within 30 days. We can deploy in a single-tenant configuration on request for regulated industries. Enterprise plans support BYO LLM (Anthropic, OpenAI, Bedrock).
How long does implementation take?
Half a day for the connector setup. Two to three days for the index to backfill (depending on the size of your Slack and Jira history). Most teams have their first usable portfolio view by end of week one.
How is this different from Glean, Atlassian Rovo, or Microsoft Copilot?
Search is table stakes. Those tools all do that well. Omnia adds the three things they don't: it provisions the project environment (Slack channels, Jira boards, Drive folders), it rolls up cost from your HRIS so you can see live burn per project, and it writes back to source systems so plans become actions.
What's the pricing model?
Per seat, per month. $19/seat on Starter, $49/seat on Team, and custom on Enterprise. No project minimum, no setup fee, and annual billing saves 20%. See pricing for what's included on each tier, or talk to us for a quote against your stack.
What happens to our data if we stop using Omnia?
You get a full export (Parquet + JSON) within 24 hours of termination. We delete all your data — primary stores, backups, and LLM caches — within 30 days, with a signed certificate of destruction. Standard data processing addendum applies.