Tia — AI Operations Assistant
Tia — AI Operations Assistant
Tia is tieback’s AI operations assistant — an intelligent layer embedded directly into your workspace for product intelligence, DPP compliance analysis, operational tracing, and governed actions.
Tia is not a chatbot or a generic AI wrapper. She is context-aware, evidence-backed, and designed to turn enterprise product data complexity into actionable operational clarity. Every response is grounded in your live workspace data, with citations and structured evidence.
Core Capabilities
Query your live workspace — products, imports, identifiers, domain configuration — and get structured, accurate operational summaries with evidence citations.
Assess field-level Digital Product Passport coverage for any product against the master DPP field catalogue. Identifies mandatory, conditional, and optional field gaps with severity-ranked recommendations. Supports nested and structured storage patterns.
Look up any passport directly by ID. Inspect snapshot identity (frozen at issuance), lifecycle state, linked entities, and detect identity drift where the passport snapshot differs from the current product master data.
Get a comprehensive operational profile of any product: full attributes, linked passports, mint batches, economic operators, compliance context, and locked-field state — all in one structured view.
Explore the DPP field catalogue structure: modules, categories, fields, value sets, and regulatory requirements explained in plain language.
Trace a product through its full operational lifecycle: batches, passports, units, status distributions, and recent mint events — end-to-end visibility.
Trace mint batches, lots, units, and events end-to-end. Investigate the full lifecycle of any operational object with break-glass override visibility.
Assess structural DPP compliance readiness: economic operators, identifiers, passport states, brand configuration, and GS1 setup — with actionable next steps.
Detect unusual patterns in mint events, audit logs, revocation trends, and actor activity to surface potential operational issues early.
Upload CSV, XLSX, PDF, or image files. Tia parses structured data, extracts text from PDFs, and analyses product labels, packaging, certificates, and scanned documents visually.
Trigger CSV manifest exports for completed mint batches through the platform’s governed export pipeline — full manifest and QR manifest formats.
Tia proposes bounded product updates with before/after diffs. You review, approve, or dismiss — nothing changes without your explicit approval. Actions execute under your identity.
Receive structured draft proposals for product attributes, import mappings, and readiness improvements — all proposals only, requiring your review.
Ask Tia about Tieback features, DPP concepts, platform workflows, and regulatory guidance. She searches Tieback’s first-party documentation and returns sourced, authoritative answers.
Resume prior conversations with full transcript rehydration. All six output types — interpretive, evidence, draft, governed-status, staged-action, and action-result — are preserved and restored.
Tia manages conversation context intelligently with safe summarisation for long sessions, ensuring coherent multi-turn interactions without context corruption.
How to Use Tia
Open Tia from the assistant panel on the right side of any tieback page. She is context-aware — she knows which page you’re on, which workspace you’re in, and tailors responses accordingly.
Ask specific, actionable questions
Tia works best with specific operational questions grounded in your data.
Resume prior conversations
Click the clock icon in the Tia header to view your conversation history. You can:
- Resume any prior conversation — the full transcript reloads with all message types intact
- Start fresh with the “New conversation” button
- Delete old conversations with a confirmation step
- See when each conversation was last active and how many messages it contains
When you resume a conversation, a small “resumed” label appears next to the context badge so you always know whether you’re continuing a prior thread or starting new.
Conversation history is private — only you can see your own conversations. History is bounded to your most recent 25 conversations per workspace. Resumed conversations send bounded prior context to the model, so Tia has awareness of what was discussed but may not recall every detail from very long sessions.
Assess DPP field coverage
Ask Tia to check how well a product’s data covers the DPP field catalogue requirements.
Example: “What DPP fields are missing for product X?”
Tia will:
- Cross-reference the product’s actual data against the active DPP field catalogue
- Show coverage percentage and severity breakdown (mandatory → conditional → optional)
- List the most important missing fields by category
- Recommend what to populate next
The DPP Field Catalogue is the master dictionary for field-level requirements. Tia uses the active published version as the source of truth. Coverage analysis checks the product’s attributes, identifiers, and top-level fields against catalogue entries, including nested and structured storage patterns where supported. Some complex storage strategies may not yet be fully covered — Tia will indicate when coverage detection is heuristic.
Inspect a passport directly
Ask Tia to look up a specific passport by its ID.
Example: “Show me passport abc-123-def”
Tia will return:
- The passport’s snapshot identity — product name, SKU, GTIN, and HS code as frozen at issuance
- Current lifecycle state (e.g., published, revoked, unpublished)
- Linked entities (mint batch, lot, unit)
- Any controlled overrides applied post-issuance
If the passport’s snapshot identity differs from the current product master data, Tia will flag the identity drift and explain that the passport preserves the identity at the time it was created.
Passport identity is a snapshot: it reflects what was true at issuance time. If the product has been updated since, the passport intentionally retains the original values. This is by design for traceability and regulatory integrity.
Explore the DPP catalogue
Ask Tia to explain what the DPP requires — modules, categories, fields, and allowed values.
Examples:
- “What modules are in the DPP catalogue?”
- “Show me the categories in the Core module”
- “What fields are in the Sustainability category?”
- “What are the allowed values for material_type?”
Tia explains catalogue structure in plain language, not raw schema dumps.
Get a product deep-dive
Ask Tia to show you everything about a specific product.
Example: “Show me everything about product X”
Tia returns:
- Full product identity and status
- Key attributes (bounded sample)
- Linked passports and their lifecycle states
- Linked mint batches
- Economic operators for the workspace
- Readiness and locked-field context
Trace a product’s lifecycle
Ask Tia to trace a product through the system.
Example: “Trace product X — show me its batches, passports, and units”
Tia traces:
- Product → Mint Batches (status, quantities)
- Product → Passports (lifecycle state distribution)
- Product → Units (status distribution, revocations)
- Recent events across linked batches
Upload files for analysis
Tia can parse and analyse uploaded files in context.
Limits: 5 MB per file, 5 files per conversation.
Trace operational objects
For complex investigations, ask Tia to trace specific objects end-to-end.
Example: “Trace mint batch abc-123 — show me everything about it”
Tia will return:
- Batch details, status, and configuration
- Associated product and lot breakdown
- Unit status distribution (minted, active, revoked)
- Recent events and any break-glass overrides
Review compliance gaps
Use gap analysis to track your DPP readiness progress over time.
Example: “What compliance gaps do I have for DPP readiness?”
Tia checks:
- Economic operator coverage (manufacturer, importer, authorised representative)
- Product identifier completeness (GTIN)
- Passport publication states
- Brand-level configuration (GS1 prefix, NACE classification)
Tia assesses structural readiness. She does not certify regulatory compliance or provide legal advice.
Trigger governed exports
Ask Tia to export CSV manifests for completed mint batches.
Example: “Export a full manifest CSV for mint batch abc-123”
Available formats:
- Full manifest CSV
- QR manifest CSV
Exports are generated through the standard governed pipeline. Tia triggers them but does not build files directly.
Review drafts carefully
When Tia identifies actionable improvements, she generates structured draft proposals — for product attributes, import mappings, or readiness summaries.
All drafts are proposals only. They require your review before any action is taken.
Approve or dismiss actions
Tia can propose bounded product updates. When she does, you’ll see a polished action card showing:
- A clear summary of what will change
- A before → after diff for each affected field
- An Approve & Execute button and a Dismiss button
- Progressive detail — expand to see all fields if more than 3 are affected
How it works:
- Nothing changes until you explicitly approve.
- Each proposal expires after 10 minutes — if you wait too long, the card shows an expired state and you can ask Tia to re-propose.
- If someone else edits the product between proposal and approval, the action is marked stale and will not execute.
- While executing, the card shows a loading state. Once complete, it transitions to a clear success or failure result.
- Dismissed actions remain visible in the transcript with a “Dismissed” label — nothing is hidden.
- Actions execute under your identity — if the backend rejects the change due to insufficient permissions, Tia will surface the error honestly.
- Currently, only users with owner or admin role can execute product updates through Tia. This is enforced by the backend, not by Tia herself. Finer-grained permission support is planned.
Example: “Update the GTIN for product X to 1234567890123”
Tia can currently propose product field updates only. Broader action types are planned for future releases. The permission model follows a “digital twin” approach — Tia proposes, the backend decides what you can actually change.
Serial Generation
Tia can generate large batches of serial numbers deterministically on the backend and deliver them as a downloadable CSV file.
Example: “Generate 5,000 alphanumeric serials with prefix TB-2026”
Tia will:
- Generate the serials using a cryptographically random, ambiguity-reduced character set
- Return a download link for a CSV file containing all serials with ordinal numbers
- Show a small sample (up to 5) so you can verify the format
- Never dump thousands of serials into the chat
Supported formats:
- Alphanumeric (default) — ambiguity-reduced charset (no I, O, 0, 1)
- Numeric — digits only
- UUID — RFC 4122 v4
Limits: Up to 50,000 serials per request. Download links expire after 1 hour.
Generated serials are not automatically inserted into any database table. They are a downloadable artifact only. To use them for minting, download the CSV and use the BYO (Bring Your Own) serials flow in the Mint Token wizard.
Ask About Tieback Features & Concepts
Tia can search Tieback’s first-party documentation to answer questions about platform features, DPP concepts, and workflows.
Examples:
- “How does passport publication work?”
- “What is identity drift?”
- “How do I generate serial numbers?”
Tia retrieves from canonical Tieback documentation and cites the source. If no relevant docs are found, she’ll say so honestly.
Coming Up in Next Release
The following capabilities are planned and in active development:
Broader write operations
Tia can currently propose product field updates through approval-gated actions. Support for creating new records, deleting records, and modifying non-product entities is in development.
Open web search
Tia can currently search Tieback’s own documentation. External web search for regulatory guidance and industry references is planned.
OCR for scanned PDFs
Text-layer PDFs are fully supported today. Scanned or image-only PDFs cannot yet be read as text. Native image files (PNG, JPG, WebP) can be analysed visually.
Support proxy mode
Cross-tenant investigation by support staff through Tia is planned. This requires additional backend infrastructure for secure proxy token handling.
Advanced statistical anomaly detection
Current anomaly analysis covers event-level patterns. Advanced statistical modelling with brand-configurable thresholds and trend analysis is planned.
Cross-conversation search
You can resume prior conversations today. Full-text search across all past conversations by keyword or topic is planned.
Privacy and Security
- Your conversations with Tia are private — only you can see them.
- Tia only accesses data within your workspace — she cannot see other tenants’ data.
- Uploaded files are processed in memory and not retained beyond the session.
- All data queries respect your role and permissions through row-level security.
- Actions execute under your identity — Tia never bypasses your permissions.
- Conversation transcripts are retained for 90 days.
- Conversation history is per-user, per-workspace — switching workspaces shows only that workspace’s conversations.
- Draft text in the composer is persisted locally per-workspace to prevent accidental loss on panel close.
- Generated serial files are stored temporarily and download links expire after 1 hour.