Standard SW workflow with CCR between CR and KO, plus three POC checkpoints for AI / Cloud capability-dependent products.
For concepts whose viability hinges on an unproven AI / Cloud capability. Time-boxed (≤2 wk) BE/model spike — no design needed.
Validate the problem statement before design proceeds.
Validate viable use cases and experience ceiling on the actual capability.
Convert the approved concept into design and engineering readiness for KO.
Approve final scope, confirm feasibility, lock cross-team commitments.
Build against approved PRD/TRD and Work Items + Dependency Log.
Validate deliverable quality and launch readiness.
Feature ships in the next available release branch.
Who does what at every stage. Highlighted cells show the function leading at that stage; striped cells indicate not directly involved.
For vertical solutions (e.g. bird cam, baby cam, garage cam) where SW is core to the value prop. HW PDP and SW workflow run in parallel and converge at four joint gates plus one joint beta program.
Form factor, sensor, optics, mounting concept, BOM ballpark.
Can the AI / Cloud capability do the job on the HW being explored?
HW team validates HW concept per HW PDP.
SW team validates problem, solution, UX — incorporating HW assumptions from the sync.
Cross-functional alignment on the integrated HW + SW vertical concept. Vertical solutions almost always meet CCR criteria (high investment, cross-team complexity).
First working prototype with target sensor / optics.
Validate AI / UX on HW POC sample, not generic dev kits. Workflow / low-fi design as needed.
Most critical sync. Locks HW spec / BOM / tooling and SW scope / design simultaneously — neither side can invalidate the other after this.
Functional defects, sensor / optics, mechanical issues. Internal / dogfood only.
Integration with new HW (driver, firmware, AI on real chip vs. dev kit), blocker bugs.
Production-representative samples. Certifications begin.
Integration / scale / reliability / billing-quota validation against DVT samples.
Single beta program: DVT / near-PVT samples + beta SW build + working subscription/billing, in the hands of the target persona (e.g. backyard birders for bird cam, new parents for baby cam).
HW manufacturing readiness + SW launch readiness as one merged decision. SW No-Go can delay HW RTM; cert slip can delay SW Go.
HW + SW + GTM all Go. For monetized verticals, subscription / billing must be live — no revenue otherwise.
Project Peek launches Wyze Day, Oct 24 2026. SW just passed ICR while HW is at EVT2 — joint checkpoints from Tab 3 were missed. This tab proposes the catch-up plan from where we are today.
60-min working session. Surface SW assumptions about HW (sensor, compute headroom, on-device vs. cloud) while HW spec still has flexibility before EVT2 gate. Output: HW spec snapshot SW commits to design against, plus any hard blockers raised to HW today.
Add HW Tech Lead + AI Lead to the CCR reviewer panel. Reframe the CCR question from "should we build this?" to "is the SW scope achievable on the now-locked HW, on the timeline that supports Oct 24?" Pre-circulate brief noting HW dependencies explicitly.
CCR approval must produce a tightly-scoped Use-Case POC charter with success criteria (accuracy thresholds, % bird species detected, latency, on-device vs. cloud cost). POC must run on EVT2 / early DVT1 samples, not dev kits.
The single most leveraged moment for hitting Jul 28 FW lock. AI team's POC results define what capabilities firmware will expose to the app. Late = FW spec slips = MP FW Lock at risk = HW airshipment math at risk. Treat as a hard milestone with explicit handoff doc.
All three tracks lock commitments together. Inputs: SW Use-Case POC outputs · FW spec · DVT1 sample availability. Output: joint Work Items + Dependency Log including FW-drop schedule for SW.
Operationalizes the "parallel FW + SW dev cycle" that closes the gap. Not a gate — a recurring sync. FW ships drops to SW weekly; SW reports back what works / what's blocked. Prevents "SW finds a capability gap after FW lock."
SW present at the FW lock decision to confirm app can deliver the vertical experience on the firmware being locked. Anything not in this firmware can only ship via OTA later — and OTA is constrained, slow. This is the line for capability commitments.
SW Beta in Aug as currently planned, but combined with HW field testing on DVT2/PVT samples. Single beta program with target persona (backyard birders), single bug log, joint triage. App can still iterate based on findings — FW cannot. Findings split into "ship in launch app" vs. "post-launch OTA."