Overview
Atlas 2.0 is a ground-up rebuild of the workspace shell: faster navigation, a shared command layer and an offline-first document engine. The goal is not a longer feature list — it is a shorter path between opening the app and doing the work that pays for it.
This brief locks the scope for the January release, names the decisions we still owe the launch team and records the assumptions we will test with the first beta cohort. Anything not listed here waits for 2.1.
Decision needed
Due FriThe public beta is holding at six weeks instead of eight. Confirm the date by Friday, or the pricing page ships without a waitlist and the launch narrative loses its strongest proof point.
Launch goals
Three numbers decide whether the rebuild worked. They are deliberately narrow: we would rather miss a vanity metric than ship a faster shell that does not change how teams work.
Activation
68%
+14 ptstarget vs. 54% baseline
Time to first doc
84s
−41%from sign-in to a saved file
Support load
0.9
−22%tickets per 100 workspaces
Timeline
The sequence is fixed around a single external date. Everything before it is negotiable; the launch week itself is not.
| Milestone | Date | Owner | Status |
|---|---|---|---|
| Feature freeze | Nov 21 | Marcus Bell | Done |
| Beta cohort | Nov 28 | Priya Nair | In progress |
| Pricing page | Dec 05 | Elena Marsh | At risk |
| Public launch | Jan 14 | Maya Okafor | Planned |
Working assumptions
- The command layer ships behind a workspace flag; no user sees it until the second beta wave.
- Legacy documents migrate in the background, with a read-only fallback for anything larger than 80 MB.
- Pricing keeps the current seat model. Any packaging change is a 2.1 conversation.
Risks & mitigations
The list is short on purpose. Each item has one owner and one observable exit condition, so launch week is about execution rather than discovery.
Migrate the 1.2M legacy documents before freeze
Marcus BellMBLoad-test the command layer at 4× peak traffic
Marcus BellMBSign off pricing copy with Legal
Elena MarshEMStaff two support rotations for launch week
Priya NairPN
Customer signal
We tested the command layer with six teams over three weeks. The consistent theme was speed of retrieval, not speed of creation: people wanted to get back to a document they had already written.
“If the command layer is as fast as the prototype, we will move every team off the old workspace. The risk is not adoption — it is going back.”
Open questions
- Do we publish the migration timeline, or wait until the first cohort clears?
- Does the pricing page need a separate waitlist for teams above 200 seats?
- Who owns the launch-week incident channel, and what is the escalation path?
Owners are listed in the review sheet. Updates land here first, then in the weekly launch note. Comments are open until Friday at 5 pm CET.