V5 CPV monitors your critical process parameters and critical quality attributes continuously, applies SPC, detects out-of-trend and rule violations in real time, and drives investigation before the batch is at risk — not after the annual review flags it.
Data pulled once a quarter; drifts and shifts identified months after they mattered.
Nelson / Western Electric rules mean different things to different analysts across sites.
'FYI: trend on line 4' — no action, no closure, no evidence of response.
Control limits, warning limits, Nelson / Western Electric rules — computed on ingestion, not on report.
Beyond fixed rules — AI surfaces gradual drifts, seasonal patterns and multi-parameter shifts that single-variable SPC misses.
Every alert becomes a CPV event with owner, SLA and closure evidence — not an email.
Threshold breaches route straight to deviation and CAPA with all context pre-linked.
The CPV report writes itself in the background; retrospective assembly is a one-click export.
CPV tools succeed when SPC is live and alerts become workflow. Criteria to test.
What it tests: Are control limits and rules computed on ingestion?
Why it matters: Quarterly is not continuous.
V5: On ingestion.
What it tests: Are Nelson / Western Electric rules applied identically across sites and analysts?
Why it matters: Inconsistency invalidates trend comparisons.
V5: Same rule engine across all sites.
What it tests: Does the tool detect drifts and shifts beyond single-variable rules?
Why it matters: Real process shifts show across parameters.
V5: AI-detected multi-variable drift alongside classical SPC.
What it tests: Do alerts become owned events with SLA and closure?
Why it matters: 'FYI' emails don't count as monitoring.
V5: Every alert is a CPV event.
What it tests: Is the CPV report a continuous artefact or a one-off build?
Why it matters: Retrospective builds are the whole problem.
V5: Report is continuously assembled.
What it tests: Do breaches route straight to deviation / CAPA with context?
Why it matters: Context lost in email is context lost in investigation.
V5: Pre-linked routing.
AI CPV vs a JMP / Minitab quarterly review and vs a legacy SPC module.
| Capability | Spreadsheet | Legacy QMS | V5 Ultimate |
|---|---|---|---|
| Cadence | Quarterly | Monthly | Live |
| Rule consistency | Analyst-dependent | Configurable | Enforced |
| Multi-variable drift | None | Rare | AI-detected |
| Alerts as records | Sometimes | Owned events with SLA | |
| Stage 3 report | Rebuild | Rebuild | Continuous |
CPV lives inside FDA Process Validation Stage 3 and its ICH / EU cousins.
An ongoing program to collect and analyze product and process data that relate to product quality must be established.
V5: Ongoing programme is the platform; monitoring is continuous, not campaign-based.
Process performance monitoring... should include an ongoing evaluation of the performance of the process.
V5: Live SPC and AI drift detection deliver the ongoing evaluation.
Ongoing process verification.
V5: V5 CPV is the EU ongoing process verification tool.
Quality risk management.
V5: CPV thresholds are set via QRM; V5 evidences the QRM at threshold-set time.
CPV goes live once CPPs and CQAs are wired to V5.
Per-product CPP / CQA list confirmed with process owners.
MES / LIMS / SCADA sources connected.
SPC rules and AI drift thresholds configured per parameter.
Alert routing to owners set; CAPA routing on for severity-level breaches.
Stage 3 report continuously assembled; retrospective is a one-click export.
CPV ROI shows in loss avoidance and inspection posture.
Live SPC + AI drift catches inside the campaign.
Continuous evidence beats retrospective assembly.
Report is a query.
One avoided lot-scale rejection typically pays the module for several years.
Setting
A sterile fill-finish site running three lines across two products.
Before
Quarterly CPV missed a gradual environmental drift; investigation started a full quarter after the drift began.
After
Six months post-cutover, a similar drift was caught inside 72 hours with a CAPA opened in-cycle; no lot was affected.
For most sites, yes — V5 SPC is IQ/OQ/PQ-supported and covers Nelson, Western Electric and custom rules. Sites with heavy JMP / Minitab investment often keep those for ad-hoc analysis and use V5 for the live layer.
Detection is deterministic given the data; the model is versioned, and detection thresholds are documented in the Stage 3 protocol. Each alert is auditable to the underlying data window.
Yes — per CPP / CQA, per site, per campaign if needed.
Free trial, no card. Live in 7 days with guided onboarding.