servicenomics

Knowledge Factory Workbench

Turn raw service material into reviewed, audience-safe answers for Cognitive Copilot, SOS, TDC, Self-Service and ALE.

IDLE V3.6
Operator mode

Inbox

Upload source material, extract evidence, and create review candidates from one place.

Next action Choose one source and follow its next step
Boundary Drafts stay internal until Review and Release & Test are complete.

Operator cockpit

Work in three areas: Inbox, Review, Release & Test. The internal five-step path stays inside the selected source.

Drafts are not production knowledge
InboxUpload material, extract evidence and create candidates.
ReviewCheck evidence, confidence, boundary and audience risk.
Release & TestRelease approved content and test the answer by audience.

Engineer/Admin Demo Guide

Guided outreach demo: service material -> Knowledge Factory -> Cognitive Copilot Evidence Trace -> SOS/TDC/ALE handoff.

0 scenes
Partner walkthrough

Show the whole value chain without pretending a draft is production knowledge.

Use this for Franke, Octonomy, b.telligent, mantro and similar conversations: the guide narrates what the operator does, what the Knowledge Factory proves and where Cognitive Copilot, SOS/TDC and ALE plug in.

No demo guide loaded yet.
No voice guide scenes loaded.

Test answer

Test what SOS, TDC, Careline, Self-Service or ALE would receive. The answer must show sources, confidence, boundary and no-answer reason when evidence is missing.

0 answer tests
No answer test yet.
No evidence trace yet.
No answer test events loaded.

Engineer/Admin: AI Agent Sessions

Parked provider-neutral Voice/Text Agent session contracts. This workbench only supports Test answer in the default flow.

Parked capability
AI Agent sessions are parked for a dedicated provider-contract slice. No session can be previewed, persisted or loaded here.
AI Agent session actions are parked. Use Test answer for active CCL validation.
No parked sessions loaded.

Operational Evidence Mining

Let Knowledge Factory infer reviewable knowledge opportunities from SOS/TDC codings, technician comments, parts usage, outcomes and follow-up evidence.

0 mining route(s)
No operational evidence mined yet.

Field Closure Feedback

Capture real TDC/SOS/FSM closure evidence and route it into Field Learning review. Evidence first, draft candidates second, no automatic truth.

0 closure item(s)
No field closure captured yet.
No closure feedback loaded.

Engineer/Admin: MCP Adapter Layer

Agent tools over governed Platform Core contracts. REST APIs remain the product boundary and source of truth.

0 tools
No MCP tool call yet.
No MCP manifest loaded.
Governed knowledge production line

Move from raw service material to approved, traceable answers without letting drafts leak into operations.

Use the workbench as a release factory: ingest sources, extract evidence, draft candidates, review with experts, publish only approved packs, then learn from field outcomes.

1 Ingest 2 Extract 3 Review 4 Release 5 Retrieve 6 Learn
Knowledge AdminQueue sources, parser jobs and artifacts.
Senior ReviewerApprove candidates, sign-offs and safety gates.
Support LeadRun retrieval, answer trace, benchmark and feedback loops.

Source Intake

TXT/MD/CSV/JSON extract now; PDF may extract coarsely or need parser/OCR.
No upload started.

Inbox

Pick one source. The selected source shows its status, readiness, audience risk and one next action.

0 sources
Needs extraction Needs review Ready to release Live Parked
No sources loaded.

Optional intake: Ticket Mining

Use when source material is a ticket batch. Paste rows, preview clusters, review patterns and promote only draft candidates.

0 batches
No ticket mining run yet.

Pattern prüfen

Screen PII and safety, then promote selected patterns into draft Knowledge Candidates.

0 patterns
No ticket batches loaded.
No ticket patterns loaded.

Causal Service Knowledge Graph

Build and trace Asset -> Symptom -> Cause -> Action -> Part -> Outcome evidence paths.

0 nodes / 0 edges
No graph action yet.
No graph evidence loaded.
No graph trace yet.

FSM Evidence Promotion Gate

Score field-derived resolution candidates before reviewer sign-off. Field learning remains evidence, not published truth.

0 gates
No field promotion gate run yet.
No promotion gates loaded.

Extraction Pipeline

TXT/MD/CSV/JSON extract immediately. PDF/DOCX/images either produce coarse text or wait for parser/OCR adapters; review section quality before drafting.

0 queued
No parser jobs queued.

Service Pattern Lab

Preview service-pattern blueprints from diagrams, BOMs, consumables and pre-market engineering context. Output stays draft/review-only.

V0.4 preview
No service-pattern preview yet.

Signal Bridge

Turn IoT, call, remote troubleshooting, field closure and partner signals into draft review candidates. This is not a predictive runtime.

V0.4.1 review bridge
No signal bridge preview yet.

Parser / OCR Worker

Create worker jobs for uploaded manuals, diagrams and scanned pages. Results stay review-only.

0 jobs
No parser locator preview yet.
No parser worker action yet.
No parser jobs loaded.

Artifact Review

Attach page/table/diagram/OCR artifacts from a parser worker and review source traceability.

0 artifacts
No artifact quality preview yet.
No parser candidate blueprint preview yet.
No candidate authoring plan yet.
No source crop manifest yet.
No structured extraction preview yet.
No structured row review plan yet.
No candidate draft preview yet.
No parser review smoke run yet.
No artifacts recorded in this session.

Candidate Review

0 candidates
No candidates loaded.
No review throughput plan yet.
No bulk review action preview yet.
No reviewer assignment preview yet.
No candidates loaded.

Snippet Editor

Select a candidate to curate it.

Visuals must be labelled as source evidence or illustration. Translations keep the same source refs, audience and safety class until separately reviewed.

No source refs loaded.

Release Candidates

0 releases
No release readiness preview yet.
No proof pack preview yet.
No export manifest preview yet.
No approval gate preview yet.
No decision memo preview yet.
No external sharing checklist yet.
No proof-room handoff bundle yet.
No proof-room runbook yet.
No releases loaded.

Published Knowledge

0 published items
No published knowledge yet.

Content Routing Matrix

Decide where curated knowledge may go after review. Runtime consumers must respect audience and safety class.

V0.2 runtime policy

Customer Remote

Only customer-safe checks, visible state, consumables, cleaning/refill guidance and escalation boundaries.

self-serviceSOS publiccustomer_safe

Careline

Guided questions, customer-safe remote steps, evidence capture and dispatch/escalation decision support.

carelineSOS agentsupervised_remote

Technician

Diagnostic steps, calibration, cleaning, replacement options, parts/tools and field evidence requirements.

TDCfield servicetechnician_only

Second Level

Deep diagnosis, repeated-fault interpretation, engineering escalation and high-risk technical boundaries.

second_levelengineeringrestricted

ALE Training

Machine-family, subsystem and role-specific training packs with source refs, images, assessment questions and gap feedback.

B2B ALElearning_designertraining pack

Visual / Translation

Source images, reviewed diagrams, generated symbolic step drawings and language variants stay tied to source and safety class.

image artifactstranslationreview required

Audience Routing Preview

Preview where the selected candidate may go before release. This explains the runtime boundary; it does not approve content.

not checked
No audience routing preview yet.
No routing decisions yet.

ALE Training Package Builder

Generate a governed B2B ALE package from published, audience-safe knowledge.

0 units
No ALE training package built yet.
No ALE learning units yet.

Cognitive Handoff Gate

Check whether a published answer can go to Customer Remote, Careline, Technician, ALE, Predictive or Partner Intelligence without leaking draft or unsafe knowledge.

not checked
No handoff gate check yet.
No runtime smoke run yet.

Runtime Production Guard

Decide whether CCL may be activated for a pilot surface. This guard requires published evidence, audience safety, no-answer behavior and explicit human operating boundaries.

not checked
No runtime production guard run yet.
No pilot activation checklist yet.
No pilot KPI baseline yet.
No pilot evidence worklist yet.
No pilot evidence intake gate yet.
No pilot evidence review memo yet.
No pilot claim boundary pack yet.
No pilot readout pack yet.

Consumer Surface Contract

Preview the exact CCL boundary for ULABSOS Customer Portal, Careline, TDC, Second Level and B2B ALE before a runtime consumer is wired.

not checked
No consumer surface contract preview yet.

BOM Compatibility Gate

Check whether a material hint may be used for an asset before CCL, SOS, TDC, CR, IBME or Replacement Engine trust it.

not checked
No BOM compatibility preview yet.

Published Retrieval

Test grounded retrieval against approved/published knowledge only.

0 matches
No embedding/index job run yet.
No retrieval run yet.
No composed answer yet.
No formulation run yet.
No matches yet.

Answer Trace / Evidence Pack

Build an auditable chain from query to retrieval, source artifacts, answer, evaluation and release boundary.

0 stages
No answer trace built yet.
No trace evidence yet.

Reviewer Sign-Off Gates

Route an answer trace to senior technician or service engineering review before release candidate readiness.

0 sign-offs
No sign-off action yet.
No sign-off queue loaded.

Release Candidate Builder

Package approved answer-trace sign-offs into a release candidate evidence pack without publishing it.

0 packages
No release candidate action yet.
No release candidate packages loaded.

Feedback Loop

Capture whether retrieved knowledge helped, or where a reviewed knowledge gap is needed.

0 feedback items
No feedback captured yet.
No feedback queue loaded.

Candidate Generation

Promote gaps and improvement signals into draft candidates and review tasks. Nothing is published here.

0 generation items
No candidate generation run yet.
No generation queue loaded.
No review tasks loaded.

Quality Evaluation

Run groundedness, citation, safety, no-answer and field-learning checks before runtime use.

0 runs
No evaluation run yet.
No evaluation scorecard loaded.

Service Resolution Benchmark

Run curated regression packs against real answer behavior before a knowledge release reaches SOS, TDC or Careline.

0 packs
No benchmark run yet.
No benchmark packs loaded.
Technical Diagnostics Optional API and troubleshooting events 
No technical events yet.