Evidence intake
Draft values with source contextExtract structured values from supplier documents
AI can help parse certificates, declarations, spec sheets, audit letters, and LCA references into proposed supplier records.
AI-assisted evidence work
AI can accelerate DPP evidence work without becoming the authority.
Redy uses AI as a governed assistant for parsing documents, drafting supplier requests, and preparing passport claims. Humans and workspace permissions still decide what gets approved, committed, and published.
AI evidence preview
source document -> proposed record -> human confirmation
Source
Supplier declaration and material certificate attached to a battery module request.
Extracted
3 fields
Status
Needs review
Authority
Supplier
Confirmation gate
Proposed values cannot update a supplier record or DPP claim until the authorized workflow confirms them.
Use cases
AI is useful where evidence work is repetitive, document-heavy, and reviewable.
Request drafting
Supplier-ready request languageTurn DPP gaps into clear supplier asks
Missing fields can become scoped requests that explain exactly what evidence is needed and why the manufacturer is asking.
Claim preparation
Reviewable DPP updatesPrepare passport claims for review
AI can assemble proposed DPP field updates from approved supplier evidence, but publication still requires governed workflow controls.
Reuse matching
Reusable evidence candidatesFind reusable evidence across requests
Suppliers can avoid rework when similar evidence already exists and permissions allow it to support another request.
Guardrails
The assistant helps with work. It does not override the workflow.
AI output is treated as proposed work until a governed Redy surface confirms it. That distinction matters for suppliers, manufacturers, auditors, and buyers.
View trust centerAI proposes. Authorized users commit.
Generated values remain drafts until an authorized supplier or manufacturer workflow confirms the action.
Scope follows the workspace lane.
Supplier workspaces and manufacturer issuer workspaces keep separate authorities, matching the portal and API model.
Source context stays attached.
Extracted values should point back to the source document, field, supplier record, and review state that produced them.
Revocation and audit are first-class.
Assistant access is designed around scoped consent, revocable sessions, confirmation, idempotency, and audit attribution.
Workflow
From source document to passport-ready evidence.
01
Supplier uploads evidence
A supplier attaches a certificate, declaration, product sheet, or other source material in the response workspace.
02
AI extracts proposed records
The assistant identifies candidate fields, values, units, and source references without treating them as approved truth.
03
User reviews and confirms
The supplier or manufacturer reviews proposed records, fixes errors, and confirms what can enter the governed workflow.
04
Approved data moves into the DPP flow
Confirmed evidence can support manufacturer review, version snapshots, API output, and passport publication.
Surfaces
AI support follows the same supplier and manufacturer lanes.
The point is not a chatbot floating above the product. It is assistant work attached to the same request, evidence, approval, and publication surfaces the team already uses.
Supplier
Respond faster without losing ownership
- Parse source documents
- Prepare reusable records
- Respond to scoped requests
Manufacturer
Review claims instead of chasing files
- Find missing fields
- Review supplier evidence
- Prepare versioned DPP updates
API / MCP
Expose assistant tools through governed surfaces
- Scoped OAuth consent
- Durable sessions
- Preview and commit semantics
Use AI where it reduces evidence work. Keep publication governed.
The AI page connects back to trust, architecture, and domain readiness so buyers can see both the workflow promise and the control model.