14-day Shadow Baseline

Measure what risky records would reach your business systems.

Measure what risky operational data would reach your CRM, ERP, analytics, WMS, or AI paths before enabling production writeback. Run Refinery in read-only or shadow mode first, then decide what should be fixed, blocked, reviewed, or verified. From €2,500 for one governed path with sample receipts and connector readiness review.

From €2,500 No production writeback required

Deliverables

• Observed record volume and issue rate.
• Issue classes: duplicate, invalid, missing, stale, conflicting, AI-unsafe.
• Auto-fixable versus ambiguous split.
• Records that would be blocked, normalized, escalated, or reviewed.
• AI-context risk summary.
• Connector readiness status.
• Sample receipts.
• Recommended first governed path.

Best first paths

Web forms to CRM, CRM to ERP, Dynamics / Dataverse customer records, Business Central customer sync, enrichment to CRM, or records feeding AI assistants.

What happens after submission

We review your path and respond within 1-2 business days. The first conversation is about scope, access mode, and whether the path is a good fit for read-only or shadow-mode measurement. We confirm connector readiness posture, agree on the minimum access required, and define what “shadow only” means for your path before observation starts.

Pricing and scope

The 14-day Shadow Baseline starts from €2,500 for one governed path. That covers scoping, read-only or shadow-mode observation, issue classification, sample receipts, connector readiness review, and a recommended first repair pilot. Additional paths, deeper ERP scope, or multi-system baselines are scoped separately after the first conversation.

The baseline is designed for teams who need proof before production writeback — not a generic free trial with hidden writeback risk. If the path is not a fit yet, we say so honestly rather than implying full connector operability.

Why shadow mode before production writeback

Most data-quality failures are invisible until they hit CRM routing, ERP sync, billing, or AI context. Shadow mode lets Refinery classify what would have happened — accepted, normalized, blocked, escalated, or reviewed — without changing production records. That reduces access scope, speeds procurement, and gives RevOps, ERP, and audit stakeholders a shared evidence base before any live repair.

Teams that skip this step often discover too late that duplicate logic, enrichment conflicts, or connector gaps make “quick fixes” unsafe. The baseline makes preventability and ambiguity visible up front.

Connector readiness honesty

Refinery separates catalog availability, setup support, readiness posture, path-binding posture, and actual runtime operability. The baseline report includes connector certification status for the agreed path. We do not claim every connector is production repair-ready. Postgres-first runtime depth is more mature today; CRM and ERP connectors may be measured in shadow mode before writeback is enabled. See the connector certification page for current posture labels.

What you get after 14 days

• Observed issue-rate summary.
• Affected records grouped by issue class.
• Deterministic versus ambiguous issue split.
• Records that would be blocked, normalized, escalated, or reviewed.
• AI-readiness risk summary for records feeding copilots, agents, or workflows.
• Recommended first governed path for a safe pilot.

Timeline

1. Day 0–2: Scope call, access mode agreement, connector readiness check.
2. Day 3–12: Shadow or read-only observation on the agreed path.
3. Day 13–14: Baseline report, sample receipts, recommended pilot scope.

Sample baseline report outline

• Path summary: source, target, record volume observed
• Issue rate and top issue classes
• Deterministic vs ambiguous split
• Records that would be blocked, normalized, escalated, or reviewed
• AI-context risk summary for copilot/RAG paths
• Connector readiness status and certification gap list
• Recommended first governed path for repair pilot

Sample receipt (illustrative)

{
  "record": "crm_account:acme-bv",
  "path": "web_form_to_crm",
  "decision": "normalized_and_escalated",
  "policy_version": "customer-master-v3",
  "writeback_status": "shadow_only",
  "issues": ["email casing", "possible duplicate"]
}

FAQ

Who is the baseline for?

RevOps, CRM owners, ERP teams, Microsoft partners post-go-live, and AI programme owners who need proof before production writeback.

What systems work?

CRM, ERP, Dynamics, Business Central, PostgreSQL, and AI-bound paths — subject to connector readiness. See connector certification.

What does it not do?

No production writeback during baseline. No claim that every connector is production repair-ready.

How is access scoped?

We request the minimum access needed to observe the agreed path — read-only API, shadow ingestion, or export-based measurement depending on connector posture. Credentials are configured through protected product surfaces after scoping, not through this marketing form.

Can we start with a Readiness Scan instead?

Yes. The free Readiness Scan is a lighter first step when you need connector and path-fit orientation before committing to a paid 14-day baseline. Many teams use Readiness Scan → Shadow Baseline → Repair Pilot.

What happens after the baseline?

Typical next steps are a Proof-Backed Repair Pilot on one governed path, Production Trust Gateway for ongoing policy-gated writeback, or pausing if the path is not yet certified for live repair. Every step should remain receipt-backed and verification-first.