Home hub · AI Engine

AI is the runtime.

Not a chatbot bolted on. Cmd-K searches every entity in your workspace; inbox triage decides what matters; cross-module signals surface anomalies before you ask.

Part of the Home hub.

What you get

Core capabilities.

Cmd-K AI search

Hit Cmd-K, type anything. Find a person, a SOP, an open task, a vendor invoice — across every hub, in one box.

Inbox triage

12 streams aggregated. AI sorts urgent / important / informational. You see a hundred items as ten.

Cross-module signals

KPI drift in Work + low kudos in Culture + slipping OKR? AI flags it as an early warning, not after the fact.

Plain-English questions

Ask: 'Which managers have falling team kudos this quarter?' Get an answer from your real data, not a generic LLM.

Reviewer copilot

Drafts review summaries from KPI + task + kudos data. Reviewers edit, don't write from blank.

Privacy-first

Your data isn't training the model. Per-workspace context, per-user permissions enforced at retrieval.

How it works

One workflow, zero context-switching.

  1. 1Cmd-K from anywhere, ask anything in plain English
  2. 2Retrieval pulls only what you have permission to see
  3. 3AI synthesizes the answer with citations to source records
  4. 4Pin useful queries as live dashboards
Common questions

Frequently asked.

Still curious? Chat with the team.

What model runs under the hood?
We use Claude 4.7 as the primary model. Some structured tasks use specialized smaller models. Per-customer enterprise pinning available on Scale.
Is my data used for training?
No. Your data is never used to train a foundation model. Per-workspace embeddings; encrypted at rest; deleted when you leave.
How does AI honor permissions?
Retrieval queries are scoped to what the requesting user can see. AI never returns content the user couldn't access via the UI.
Does AI cost extra?
Capped on Growth; unlimited on Scale. Most teams stay under the cap; we tell you when you're approaching.

Stop juggling tools.
Start running the business.