CJ
What are the model risk management requirements under SR 11-7 for our credit scoring model, and which apply before we can promote CS-204 to production?
neurology
AveryVirtual expert
SR 11-7 frames model risk management around three pillars: robust model development, ongoing validation, and governance with clear accountability. For CS-204, the controls that gate production sit in the first two pillars.
An independent validation covering conceptual soundness, outcomes analysis, and monitoring design is required prior to deployment, performed by staff independent of model development. Development documentation must let a third party reproduce the model without recourse to the developers.
Because CS-204 informs consumer credit decisions, fair lending testing and a documented effective challenge record should be attached to the validation package. Your MRM policy also requires Tier-1 models to carry a board approved risk rating, which CS-204 currently lacks.
Summary
Bottom line: CS-204 cannot be promoted until independent validation is complete and a board approved Tier-1 rating is issued.
- check_circleIndependent validation is mandatory pre-production.
- check_circleReproducible development documentation must accompany the package.
- check_circleFair lending testing required for consumer-credit use.
- check_circleOpen blocker: no board approved risk rating on file.
Bullets
- SR 11-7 three-pillar framework
- Development, validation, governance
- Pre-production gates for CS-204
- Independent validation, reproducible documentation
- Consumer-credit considerations
- Fair lending testing, effective challenge
- Open blocker: Tier-1 board rating
account_treePremise chain4 premisesexpand_more
CS-204 is classified a Tier-1 model under MRM policyT1
Independent validation is required before productionT1
Tier-1 models require a board-approved risk ratingT2
CS-204 has no board-approved rating on fileT1
grid_onRisk register rowfields mapped from this answer
infoEach field is mapped and editable before export. New templates are added under Take action → Manage templates.
CJ
Who has to sign off the validation, and can development staff do it?
neurology
AveryVirtual expert
Sign-off must come from a validation function that is independent of model development. Developers may support the exercise but cannot own the validation conclusion or the effective-challenge record.
gavel
SR 11-7: Guidance on Model Risk Management
gavelFederal Reserve · Regulation · pp. 3–7
…validation should be performed by staff who are independent of model development, covering conceptual soundness, ongoing monitoring, and outcomes analysis…
Why relevant: directly defines the independent-validation requirement and who may perform it.
rule
Independent Validation Standard MRM-STD-02
ruleInternal · Policy · §4.2
…the scope of an independent validation shall include benchmarking, replication where feasible, and an assessment of requirements for ongoing monitoring…
Why relevant: your internal standard operationalising the regulator's expectation.
description
OCC Bulletin 2011-12
descriptionOCC · Regulation · supervisory
…supervisory expectations for validation activities, including effective challenge by parties independent of the model owners…
Why relevant: companion supervisory guidance reinforcing effective challenge.
menu_book
Outcomes Analysis Playbook
menu_bookInternal · Method · v2
…operational guidance for the validation team performing outcomes analysis and back-testing…
Why relevant: method-level detail, useful but not authoritative on requirements.
CJ
Map EU AI Act Articles 9–15 obligations to our internal model-risk controls and flag any requirements we do not yet cover.
travel_explore
AveryDeep research
helpClarified before starting
Which articles are in scope?Art. 9–15 (risk, data, documentation, transparency, human oversight)
What output do you want?A control-mapping table plus a list of gaps
How deep, and which sources?Standard depth · prefer T1 and T2 · internal policy first
account_treeReasoning plancaptured and preserved with this inquiry
check1 · Plan. Broke the objective into seven sub-questions, one per article cluster.0:08
check2 · Initial search. Retrieved the EU AI Act text and 38 internal control and policy sources.0:51
check3 · Reflect. Assessed coverage and trust tier; two clusters were thin on internal evidence.1:20
check4 · Follow-up. Ran four targeted searches for human-oversight and record-keeping controls.2:05
check5 · Synthesise. Aligned each obligation to a control and flagged the unmapped ones.2:48
progress_activity6 · Improve. Running a sub-plan to confirm the two gaps are not covered elsewhere.71%
7 · Choose format. Mapping table plus gap list, as requested.queued
8 · Produce output. Render the report and issue a validity warrant.queued
Draft, updating as the run completes. Articles 9–15 map cleanly onto existing model-risk controls in most areas. Two obligations have no current internal control and are flagged below.
| EU AI Act obligation | Internal control | Status |
|---|---|---|
| Art. 9 · Risk management system | MRM-STD-01 model risk framework | Mapped |
| Art. 10 · Data and data governance | DATA-GOV-03 data quality and lineage | Mapped |
| Art. 11 · Technical documentation | CS-204 development record v4 | Mapped |
| Art. 12 · Record-keeping and logging | None | Gap |
| Art. 13 · Transparency to deployers | MODEL-CARD-02 disclosure pack | Mapped |
| Art. 14 · Human oversight | None | Gap |
| Art. 15 · Accuracy and robustness | MRM-STD-02 validation standard | Mapped |
Two gaps to close. Article 12 needs a defined automatic logging and retention control for the model's decisions, and Article 14 needs a documented human-oversight procedure for the credit-decisioning use of CS-204.
Summary
Bottom line: five of seven obligations under Articles 9 to 15 are already covered; two require new controls before deployment.
- check_circleMapped: risk management, data governance, documentation, transparency, accuracy.
- check_circleGap: Article 12 logging and retention control.
- check_circleGap: Article 14 human-oversight procedure.
Bullets
- Coverage of Articles 9 to 15
- Five obligations mapped to existing controls
- Gaps
- Art. 12 logging, Art. 14 human oversight
- Recommended next steps
- Draft two controls, re-run mapping
Risk & Compliance basechevron_right5W1Hchevron_rightIndependent validation
category5W1H
descriptioncheck_circleIndexedgroupsSharedexpand_more
CS-204_Dev_Record_v4.pdf
Development record for the CS-204 credit-scoring model: data, methodology, performance testing, and limitations.
Type Procedure auto · changeOrigin UploadedTier T2Visibility Shared
rulecheck_circleIndexedhow_to_regNominatedexpand_more
Fair_Lending_Protocol_draft.docx
Draft protocol for fair-lending testing of consumer-credit models, including disparate-impact analysis.
Type Policy auto · changeOrigin CrawledTier T1Status Nominated
datasetprogress_activityProcessinglockPrivateexpand_more
Validation_backlog_Q2.xlsx
Quarterly tracker of models awaiting validation, with owners, due dates, and tiering.
Type Reference auto · changeOrigin CrawledTier T3Visibility Private
grid_oncheck_circleReusablegroupsSharedexpand_more
Risk_register_row.tmpl
Reusable risk-register row mapped from an answer; fields editable before export.
Type TemplateOrigin SavedStatus ReusableVisibility Shared
cloud_sync
Local Drive · /Compliance
running · 1,204 discovered · 892 indexed · 312 queued
ruleInclude: *.pdf, *.docxblockExclude: /archivelayersDepth 3 · 5k pagesgroupsShare → CompliancetuneSelection criteria
how_to_reg14 nominated for approval
public
Regulator sites · scheduled
paused · next run daily 06:00 · recursion 2
listSeeds: 6 URLsaccount_treeTopics: Model Risk, Basel IVgroupsShare → all collections
hard_drive
Local Drive · /Compliance
892 files · last sync 4m ago · auto-crawl on
cloud
SharePoint · Risk
connected · scoped to /Risk/Policies
Drives are added in one place: the Connect a drive button above. Scope, schedule, and sharing are set per drive.
Prudential
AI & Models
Suggested for you
addOperational ResilienceaddClimate Risk DisclosureaddCCAR / Stress Testing
auto_awesomeDiscover topics to follow
Describe a subject in your own words. The base matches it to topics, tags, concepts, and 5W1H aspects, then suggests what to follow and how well it is already covered.
search
Tryclimate risk disclosuremodel validation independenceGenAI governance
By tag
By concept
By aspect · 5W1H
By framework
Suggested for you
The projection is relative to the current context. Drill into a node to focus it; the breadcrumb tracks where you are. Switch the projection type above.
EUR
EU AI Act T1 source
Delegated act on high-risk AI classification published
European Commission·2h ago
Abstract: The Commission sets the criteria and conformity steps for high-risk AI systems, tightening how credit-decisioning models are classified and what evidence must accompany them before deployment.
Why relevant: affects how CS-204 is classified; 4 sources in AI Governance may need re-validation.
Suggested inquiries
question_answerDoes this reclassify CS-204?question_answerWhich controls need re-validation?
OCC
Model Risk T1 source
OCC signals refresh of model risk management expectations
OCC·Yesterday
Abstract: The OCC flags a forthcoming update to supervisory expectations for model risk, with emphasis on validation independence and ongoing monitoring cadence.
Why relevant: touches your most-cited branch; may change validation cadence.
IND
Model Risk T3 source
Industry blog: practitioners debate validation independence
RiskBlog·2 days ago
Abstract: A practitioner roundup of how firms staff and scope independent validation, with varied interpretations of who may sign off.
Why relevant: contextual only, not authoritative. Use Open case to Steward to request this source be vetted.
Dismissed stories are kept in When → Dismissed and on the source's history, so a source is never silently dropped.
CASE-204 · Fair lending source disputed
hourglass_topTriageSource dispute·2m ago
CJ flagged that the fair-lending source cited in INQ-204 may be superseded by a newer protocol. Awaiting steward triage and a source re-check.
linkINQ-204descriptionFair Lending Protocol
CASE-198 · Benchmarking gap escalation
errorOpenData gap·1h ago
The benchmarking methodology rests on a single unverified source with no corroboration. Escalated so a vetted source can be added before the analysis is relied on.
linkINQ-191
CASE-191 · GDPR Art. 22 interpretation
check_circleResolvedHelp request·3 days ago
Asked which framework crosswalk covers automated decisioning. Answered with the GDPR crosswalk, Article 22, and the linked guidance. Resolved and recorded.
linkINQ-203
priority_highNeeds you6 itemsagent-surfaced problems · fix first
ConflictProblem detection · 1 new · 1h
4 conflicts, 1 new: SR 11-7 (T1) against a decaying T3 note on validation sign-off.
CollectionCollection run · 6h
Intranet crawl RUN-307 on /FinCrime/Policies failed at the Fetch step: authentication expired.
RecommendationsRecommendation agent · today
7 recommendations await your review, preview-first; commit them as a batch.
Decaying premisesMaintenance agent · oldest 94d
2 premises are decaying: the peer-reviewer sign-off note (94d) and an unreachable vendor methodology.
GapsProblem detection · 9 candidates ready
23 open gaps; 9 pre-staged candidates are ready to assign.
CasesFrom users · 2 new · 2m
3 open cases, 2 new: a source request and an answer dispute.
Turn the problem queue into a document
Recommendations
7
recommendawaiting you
Conflicts
4
north_east1 new
Coverage
87%
trending_up+4.2 pts
Freshness
92%
schedulecurrent
expand_moreMore metricsknowledge items · active users · queries · open cases
What's in the baseexpand_more
By origin and by knowledge family · 18,420 items
By origin
Subscriptionlicensed packs
8,470
Organizationalyour content
7,020
Internetvetted external
2,930
By family
Regulation & rules
5,510
Policy & standard
4,420
Procedure & guide
3,690
Reference & data
2,950
Memo & note
1,850
Casesexpand_more
Raised to you by users
Subscription updatesexpand_more
Licensed packs and consolidated bases
Agentsexpand_more
Loops you direct, functional view · machinery lives under Admin
travel_exploreDiscoveryautorenewrunning
troubleshootProblem detectionautorenewrunning
recommendRecommendationhow_to_regawaiting you
monitor_heartMaintenanceautorenewrunning
Recommendationsexpand_more
Top items the agents surfaced for you to act on
GapBenchmarking
Pre-staged candidate: BIS benchmarking note (T1).
ConflictValidation sign-off
SR 11-7 (T1) vs a decaying T3 note. Untrust the T3 premise to resolve.
Recommendations
7
recommendawaiting you
Open gaps
23
north_east2 new
Pre-staged candidates
9
inventory_2ready to admit
Conflicts
4
north_east1 new
Loading the queue…
task_alt
Nothing waiting
No recommendations, gaps, or conflicts need you right now. The agents keep watching, and new items will appear here.
error
The queue did not load
Something went wrong loading the curation queue. This is usually temporary. Retry, and if it persists, check the platform status.
Coverage heat map
Cell area is the selected measure, colour is coverage depth · click to drill in, click a gap to stage research · measure and whitespace under Options
All topics
thinmiddeepgap
Coverage matrix
Branch by dimension, colour is coverage strength · click a cell to drill into Contents
ThinStrong
Knowledge graph
A windowed ego view: a centre claim and its neighbours to a chosen hop radius · click a node to re-centre, click the centre to inspect · hops, mode, scope, and colour under Options
T1T2T3conflict
Topic map
The subject graph as a windowed ego view: a centre topic and its neighbours · colour is coverage depth · click to re-centre, click the centre to open in Contents · hops, mode, scope, and colour under Options
deepmidthingap
Validity landscape
Validity over time with decay · move the as-of point to see what the base knew on a date
The as-of point moves from Options; the context strip above records it.
Decaying premises, refresh due
Topographic
Height is epistemic gravity, colour is trust, basins are thin or decaying areas
T1 peakT2 peakthin basin
Coverage radar
Declared, observed, and available, per dimension
Observed Available Declared
Capital & LiquidityBasel III / IV
94%
Model Risk MgmtSR 11-7 / OCC
90%
Data PrivacyGDPR / CCPA
88%
Financial CrimeAML / KYC
72%
AI GovernanceEU AI Act
61%
Framed reading
The selected branch read through the chosen frame · the same engine as the user Canvas, seeded from scope · pick the branch and frame in the drawer
category5W1H
account_treeInspect
Trace the premises behind a node, cell, or region
searchsend
Sign-off must come from a validation function independent of model development. Developers may support the work but cannot own the conclusion.
account_treePremise chain
Validation must be independent of developmentT1
Developers may not sign the conclusionT1
A peer reviewer may sign in small teamsT3
Drawn from 4 sources
Beliefs
Items
18,420
trending_up+312 (30d)
Subscription
8,470
workspace_premium46%
Organizational
7,020
corporate_fare38%
Internet
2,930
public16%
Items
Click an item to open the viewer · personal vault content is never shown here
Licensed packs
4
workspace_premium8,470 items
Updates available
2
north_eastreview
Consolidated bases
1
account_tree2 members
Pack freshness
96%
schedulecurrent
Update pipeline
One status per pack and environment, read the same by the admin, the partner, and you: published, available, applied · the admin makes a version available, you review the changelog and apply
DEV applies automatically on release. On UAT and PROD an available version waits for your review: the content changelog and the recombination impact show before anything commits.
Licensed packs
Updates recombine into your base automatically without disturbing local content · apply is preview-first
Consolidated knowledge bases
Composites that query several bases together, each base warranted separately
Members must come from one environment. Each base is queried and warranted separately; results combine by the chosen strategy with a conflict-resolution policy.
Active users (7d)
142
group+11
Queries (7d)
1,084
trending_upsteady
Vaults connected
38
inventory_22 nominating
Open cases
3
contact_support2 new
Peopleexpand_more
Activity by user, role-scoped · no content shown
Vault visibilityexpand_more
Personal vaults: counts and sync only, never the content
A steward sees that a vault exists and its counts, to manage nominations to the shared base. The files themselves stay private to their owner.
Casesexpand_more3 open
Raised to you by users
Activity logexpand_more
Warranted changes to the base, newest first · every entry links to its warrant
Top queriesexpand_more
Aggregate topics asked · no content shown
Model validation
284
Capital adequacy
196
GDPR / privacy
142
EU AI Act
118
AML / KYC
86
Open
3
north_east2 new
In triage
1
hourglass_topwith you
Resolved (30d)
14
check_circle+5
Median response
3h
trending_down-1h
Loops running
4/4
bolthealthy
Candidates found (7d)
12
trending_updiscovery
Awaiting you
7
how_to_regproposals
Conflicts open
4
error1 new
DiscoveryFind candidate sources for declared gapsautorenewRunning
23 declared gaps · 12 candidates (7d) · proposes, you approveexpand_more
watching23 declared gaps
found (7d)12 candidates
autonomyproposes · you approve
loop
Problem detectionFind gaps, conflicts, and decayautorenewRunning
23 gaps · 4 conflicts · 21 revaluing · proposes, you approveexpand_more
found23 gaps · 4 conflicts
decay21 items revaluing
autonomyproposes · you approve
loop
RecommendationStage candidates against the gapshow_to_regAwaiting you
7 proposals staged · 18 accepted (30d) · proposes, you approveexpand_more
staged7 proposals
accepted (30d)18
autonomyproposes · you approve
loop
MaintenanceRevalidate and keep sources freshautorenewRunning
21 revalidations · freshness 92% · auto within your scopeexpand_more
revalidations21 in progress
freshness92%
autonomyauto · within your scope
loop
Sources
503
trending_up+18
T1 / T2
412
shield82%
Stale
21
schedulerefresh due
Unreachable
3
link_offre-source
search_off
No sources match
No sources match the current search and filters. Clear them to see the full set, or add a source.
error
Sources did not load
The source list could not be loaded. This is usually temporary. Retry, and if it persists, check the platform status.
Recipe as coderecipe.yaml
Declarative and version-controlled
scope: domain: Risk & Compliance dimensions: [geography, framework, time, sector, authority, topic] sources: - id: fed-sr-11-7 tier: T1 monitor: weekly - id: occ-mrm tier: T1 monitor: weekly collection: web: managed discovery intranet: managed crawler exclude: [/archive, opinion-blogs] governance: veracity_gate: 0.7 pii_detect: on_ingestion pii_default: redact pii_purposes: [kyc, onboarding] warrant: on_answer
Components
Navigate and edit the same recipe
Version history
Each revision is a commit; roll back if needed
Promotion
Move this recipe along the environment path
DEV · v14arrow_forwardUAT · v14arrow_forwardPROD · v13
Promotion is preview-first: a recipe diff is shown, approval is required, and the apply writes a warranted audit entry. Environments are managed under Admin.
priority_highNeeds attention6 itemsinfrastructure and activity · fix first
Service degradedModel endpoint · failover active
Model endpoint is degraded: one of two pools is slow, p95 980 ms. Failover is carrying queries.
Jobs2 failed or stuck · 24h
2 jobs need a look: warrant-chain-checkpoint retried once at :00, and ingest batch ING-2214 is stuck since 06:00.
Base on watchMarkets & Treasury · PROD
Markets & Treasury is on watch: p95 2.6 s, above target while the model endpoint is degraded.
Budgetbreaker at 75%
Cascade depth is at 3 of 4, near its circuit breaker; query overage is $182 MTD.
Keys2 expiring
2 keys approach rotation: webhook-signing in 6 days, github-deploy-key in 10 days.
Access1 anomaly · today
copilot-token-04 is over its daily bound: 520 of 500 queries. Overage applies; the auditor sees the same signal.
API uptime
99.98%
check_circle30 days
p95 query latency
1.8s
trending_down-0.3
Error rate
0.20%
check_circlenominal
Queries / min
640
trending_uppeak hour
expand_moreMore metricsingest throughput · claims written today
Service health
click a service for configuration and detailchecked 0s ago
APIopen_in_full
uptime99.98%
p95120ms
MCP serveropen_in_full
uptime99.97%
p95130ms
Query serviceopen_in_full
uptime99.95%
p951.8s
Ingestionopen_in_full
uptime99.90%
queue312
Knowledge storeopen_in_full
uptime99.99%
p958ms
Model endpointopen_in_full
statusDegraded
p95980ms
Discovery planecloud
uptime99.8%
scopeexternal
Environments & knowledge bases
live status per base · metrics every 5s
PRODproduction · recipe v13check_circleHealthy
Risk & Compliance
expand_morePRODcheck_circleHealthy
Queries / min
240
p95 latency
1.7s
Claims today
1,284
Freshness
92%
Activity
live
Answer pipeline
Workers
Markets & Treasury
expand_morePRODwarningWatch
Queries / min
120
p95 latency
2.6s
trending_upabove target
Claims today
540
Freshness
88%
Activity
live
Answer pipeline
Workers
UATacceptance · recipe v14check_circleHealthy
Risk & Compliance
UATcheck_circleHealthy
Queries / min
60
p95 latency
1.6s
Claims today
1,190
Freshness
95%
Activity
live
Answer pipeline
Workers
Markets & Treasury
UATcheck_circleHealthy
Queries / min
34
p95 latency
1.5s
Claims today
505
Freshness
93%
Activity
live
Answer pipeline
Workers
DEVdevelopment · recipe v14buildBuilding
Sandbox
DEVbuildBuilding
Queries / min
8
p95 latency
1.4s
Claims today
110
trending_upbuilding
Freshness
60%
Activity
live
Answer pipeline
Workers
Subscribed experts
3
workspace_premiumactive
Upgrades available
2
upgradeto review
Auto-apply
DEV
bolton release
Renewal
Mar 2027
eventannual
Subscribed virtual experts
Each domain expert ships its licensed content and a recommended recipe; versions are applied per environment
Risk & Compliance expertlicensed content + recipeupgrade v2025.7
currentv2025.6
DEVv2025.7
UAT / PRODv2025.6
domain expert
Markets & Treasury expertlicensed content + recipeupgrade v2025.5
currentv2025.4
DEVv2025.5
UAT / PRODv2025.4
domain expert
Custom expertincluded with base licenceup to date
currentin-house
DEV / UAT / PRODaligned
custom
Update pipeline
One status per pack and environment, read the same by the admin, the steward, and the partner: published, available, applied
The admin sees the technical envelope only: version, size, compatibility, and the schedule window. The content changelog is the steward's to review at apply time.
Upgrade application policy
How a new expert version moves through each environment: DEV automatic; UAT and PROD admin makes available, steward applies
Applying an expert upgrade promotes its recipe and content; the apply is preview-first and warranted, and runs through Migration. The base stays queryable while it applies. On UAT and PROD the admin stages availability and the steward applies after reviewing the content changelog.
Pending promotions
1
north_eastUAT → PROD
Recipe (main)
v14
commitCI passing
Apply in progress
1
syncDEV re-embed
Drift
0
check_circlein sync
Promotion pipeline
Move a validated recipe one stage along the path; each promotion shows a diff and requires approval
DEV · v14arrow_forward
UAT · v14arrow_forward
PROD · v13
Promotions are preview-first: the recipe diff is shown, approval is required, and the apply writes a warranted audit entry. Rollback returns the environment to the previous tag.
Recipe state by environment
Deployed version, what is available to apply, and the controls to move it
Applying a recipe can trigger re-embedding or index migrations; those run as background jobs under Operations, and the base stays queryable while they run.
Migration history
Every promotion and apply, recorded
Recipe as coderecipe.yaml
Declarative and version-controlled
scope: domain: Risk & Compliance dimensions: [geography, framework, time, sector, authority, topic] sources: - id: fed-sr-11-7 tier: T1 monitor: weekly collection: web: managed discovery intranet: managed crawler exclude: [/archive, opinion-blogs] governance: veracity_gate: 0.7 pii_detect: on_ingestion warrant: on_answer
Components
Navigate and edit the same recipe
Recipe per knowledge base
Each base has its own recipe and deployed version; promote on the Migration tab
Risk & Compliancev13 · PROD
draftv14 ready
CIpassing
recipe
Markets & Treasuryv13 · PROD
draftv13 current
CIpassing
recipe
Sandboxv14 · DEV
draftediting
CIpassing
recipe
Active agents
14/16
bolthealthy
Queue depth
312
trending_down-44
Ingest rate
1.2k/min
trending_upsteady
Errors (1h)
3
trending_downcontained
priority_highNeeds attention2 itemsloops needing a decision
Awaiting approvalreview-inconsistency · L0
review-inconsistency holds 4 proposals at L0, propose only. Nothing writes until a person approves.
Approval queue3 items · coordinator
3 queued actions across the loops wait on an approval decision; the oldest is 2 hours.
discovery-webCandidate sources · scales outautorenewRunning
instances×3
autonomyL1 · approve writes
throughput48 found
discovery
ingest-webCollect and parse accepted sources · scales outautorenewRunning
instances×4
autonomyL1 · approve writes
throughput126 in flight
ingestion
process-embedChunk, type, embed, write · scales outautorenewRunning
instances×6
autonomyL2 · auto
throughput1.2 k/min
processing
crawler-localVault drive ingestionautorenewRunning
instances×1
autonomyL2 · auto
throughput312 queued
ingestion
coordinatorLoop oversightautorenewRunning
instancessingleton
autonomysupervisor
budgetn/a
watching6 loops
oversight
review-inconsistencyConflicts and decayhow_to_regAwaiting approval
instances×1
autonomyL0 · propose only
queue4 proposals
review
Running jobs
7
bolthealthy
Scheduled
12
schedulenext 06:00
Failed (24h)
2
trending_downretried
Retention
365d
lock_clockwarrants ∞
priority_highNeeds attention2 itemsfailed or stuck · 24h
Retriedwarrant-chain-checkpoint · hourly
warrant-chain-checkpoint failed at :00 and was retried once, then OK. Watch the next run.
StuckING-2214 · since 06:00
Ingest batch ING-2214 is stuck at the parse step since 06:00; retrying with backoff, 14 items held.
Regulator site monitorcron · daily 06:00check_circleOK
last runtoday 06:00
next runtomorrow 06:00
cron
Decay revaluationcron · nightlycheck_circleOK
last run02:00
next runtonight 02:00
cron
Vault drive syncevent · on changeautorenewRunning
last event4m ago
synced today38 files
event
Warrant chain checkpointcron · hourlyerrorRetried
last run:00
retries1 then OK
cron
Charges MTD
$3,140
paymentssubscription + usage
Overage MTD
$214
trending_upqueries + warrants
Named seats
96/120
badgeassigned / licensed
Breakers tripped
0
check_circlenominal
priority_highNeeds attention2 itemsbudgets near a breaker
Breakercascade depth · 75%
Cascade depth is at 3 of 4. One more level trips the breaker and pauses the pool, loudly.
Overage$214 MTD · climbing
Metered overage is climbing: $182 in queries and $32 in produced warrants this month.
Charges this period
Base subscription, content subscriptions, and metered overage
Base licence$25,000/yr
Content subscriptionsincl. experts
Additional knowledge base$10,000/yr
Query overage MTD$182
Warrant overage MTD$32
Circuit breakers
Trip loudly and pause the pool
Daily spend$48 of $120
40%
Retractions6 of 25
24%
Cascade depth3 of 4
75%
Error ratecontained
12%
Max steps / run120 of 200
60%
No-progress timeout90s cap
ok
Autonomy & budgets
Authority per action class, and token budget per run and per day
ProposeL2 auto
Write to baseL1 approve
Retract / demoteL0 propose
Per-agent token budget
discovery-webL1
crawler-localL2
review-inconsistencyL0
Users
86
trending_up+4
Stewards
7
account_tree2 bases
Agent tokens
11
keyscoped
Pending
2
hourglass_emptyinvites
priority_highNeeds attention2 itemsanomalies first
Anomalycopilot-token-04 · today
copilot-token-04 is over its daily bound: 520 of 500 queries. The scope is read-only; overage applies.
Invites2 pending · oldest 6d
2 invites are pending; j.okafor@ has been waiting 6 days.
Chris Jacksonperson · SSOcheck_circleActive
baseRisk & Compliance
personaUser + Steward
PII modeRedact
person
A. Reyesperson · SSOcheck_circleActive
baseRisk & Compliance
personaAuditor
PII modeBlock
person
copilot-token-04agent · scoped tokencheck_circleActive
baseRisk & Compliance
personaExplore (read)
PII modeRedact
agent
j.okafor@person · invitedhourglass_emptyPending
baseMarkets & Treasury
personaUser
PII modeRedact
person
Agent tokens
Issue, rotate, and revoke scoped tokens; each carries an authority matrix
copilot-token-04
ingest-bot-02
Agent repository grants
Explicit, least-privilege, per repository, revocable; every fetch is logged
SharePoint · Policies
File share · Risk
databasecheck_circleHealthy
Risk & CompliancePROD
503 sources · 1,284 claims · 86 users
Isolation: sharedResidency: asia-southeast1Retention: 365d
databasecheck_circleHealthy
Markets & TreasuryPROD
214 sources · 540 claims · 31 users
Isolation: dedicated storeResidency: me-central1Retention: 365d
sciencebuildBuilding
SandboxDEV
42 sources · 110 claims · 5 users
Isolation: sharedResidency: asia-southeast1Retention: 30d
Environments
3
lanDEV · UAT · PROD
Recipe version
v14
commitmain
Pending promotions
1
north_eastUAT → PROD
Drift
0
check_circlein sync
Environments
Recipe version deployed in each
DEVdevelopmentcheck_circleCurrent
recipev14 · main
promotes fromsource
promotes toUAT
basesRisk & Compliance, Sandbox
dev
UATacceptancecheck_circleCurrent
recipev14 · release/14
promotes fromDEV
promotes toPROD
basesRisk & Compliance, Markets & Treasury
uat
PRODproductionnorth_east1 behind
recipev13 · tag v13
promotes fromUAT
promotes toproduction
basesRisk & Compliance, Markets & Treasury
prod
Code repository & promotion keys
The GitHub repository and the keys used to move recipes between environments
hubgithub.com/acme/corvair-recipesmain
commitv14: add EU AI Act monitoring + tier rules2h ago
check_circleCI: recipe validated, 0 errorsverifiedPassing
Promotion keys · status only, never values
Promotion path
Each environment promotes to the next; movement and application are controlled on the Migration tab
DEV · v14arrow_forward
UAT · v14arrow_forward
PROD · v13 · 1 behind
Promotions are preview-first: the recipe diff is shown, approval is required, and the apply writes a warranted audit entry. Rollback returns the environment to the previous tag.
Last backup
2h ago
check_circleverified
Backup retention
35d
lock_clockdaily
Embedding model
v3 · 1536d
boltcurrent
Migrations
0
check_circleup to date
Backups & restore
Point-in-time snapshots; restore is preview-first and recorded
Re-embedding
Run when the embedding model or dimension changes; the base stays queryable
Current modelactive
Last re-embedcheck_circlecomplete
Migrations
Read-only summary. Movement and application of recipes and schema migrations are controlled on the Migration tab.
Schema and index migrationscheck_circleUp to date
Discovery plane
The outward search service that finds candidate sources; managed by connection, policy, budget, and health
A one-way governed seam pushes policy outward and pulls only candidate finds back; ingestion in the knowledge plane decides what is accepted. The grounded-search provider sits behind the seam and is not exposed to tenants.
Active users today
78/96
groupof assigned seats
Queries today
38,200
bolt500 / seat / day
Warrants produced
1,240
verified_user20 / seat incl.
Users over limit
3
trending_upoverage applies
Per-user usage today
Each seat includes 500 queries and 20 produced warrants a day; overage is metered
Query usage
Each seat includes 500 queries a day; overage is metered per query
Risk & Complianceavg 410 / 500 per seat
82%
Markets & Treasuryavg 180 / 500 per seat
36%
Overage queries today$182
Validity warrant production
A warrant is produced from the logs on first view or export, then cached and viewed any number of times at no further charge. 20 produced per seat per day are included.
Produced today1,240 of ~1,920 included
65%
Served from cacheno charge
8.6k
Overage warrants today$32
Cache hit ratecheck_circleefficient
Secrets
9
lockmanaged
Due for rotation
1
schedulein 6 days
Signing key
KMS
verified_usergovernance-held
Static keys
0
check_circleworkload identity
priority_highNeeds attention2 itemsexpiring keys first
Due soonwebhook-signing · 6 days
webhook-signing is 84 days into a 90-day policy. Rotate before it lapses; the API import path depends on it.
Due soongithub-deploy-key · 10 days
github-deploy-key is 20 days into a 30-day policy. Recipe pulls on apply use this key.
Managed secrets
Held in the secret manager; values are never shown
graph-db-credentialslockManaged
used byquery, ingestion
last rotated42 days ago
policy90 days
secret
model-api-accesslockManaged
used byembed, generate
last rotated12 days ago
policy90 days
secret
webhook-signingscheduleDue soon
used byAPI import
last rotated84 days ago
policy90 days
secret
Warrant-signing key
Held by the governance identity, separate from the application accounts
keywarrant-signing-key (KMS)HSM-backed
scheduleRotation schedule: every 180 dayslast 60 days ago
verified_userHeld by the governance identity, not the app accountscheckseparated
Warrants (30d)
1,942
trending_up+128
Verified
1,942
check_circle100%
Failed verify
0
verified_userchain intact
Exported
37
outboxto reviewers
link
Hash chain verified, 1,942 of 1,942 links intact
Each warrant seals the previous warrant's hash. A break anywhere would surface here.
a91flink7c0elinkd4b2linklive
Warrants are read-only here. The auditor verifies the chain and exports with selective disclosure. Each warrant is issued by the system at the moment of an answer, decision, or access, and cannot be altered after the fact.
Requirements
42
checklistSR 11-7
Covered
36
trending_up86%
Disclosure gaps
6
north_east2 high
Last reviewed
3d
historyA. Reyes
Coverage by section
Where the gaps sit in SR 11-7
Development & implementation
95%
Validation
88%
Governance & controls
82%
Ongoing monitoring
71%
A gap is a disclosure the base cannot yet warrant. The auditor flags it as a case to the steward, who owns the fix. The auditor does not curate the evidence.
Conflicts
4
north_east1 new
Unresolved
3
pendingin triage
Single-source
9
linkdependencies
Decaying
12
trending_downpremises
Conflict mapdownload
Nodes are admitted sources; the red edge is the contested claim
T1 authoritativeT2 reliableT3 contextualT4 unverified
Open conflicts
Each can be inspected in its warrant or referred to the steward
ContradictionModel Risk · Validation
SR 11-7 (T1) requires independent validation; an internal working note (T3) permits a peer reviewer in small teams. The T3 premise is decaying.
Tier inversionINQ-176 · answer
An answer leaned on a T3 industry blog above a T1 supervisory source on validation independence. The inversion is flagged for review.
Single-sourceBenchmarking
Benchmarking methodology rests on a single unverified source. No corroboration.
SupersededINQ-204 · fair lending
A fair-lending source cited in INQ-204 may be superseded by a newer protocol. Currency is in doubt.
Decayingocc.gov · 94d
A load-bearing premise depends on a source last validated 94 days ago.
High validity
71%
trending_upstable
At risk
17%
scheduledecaying
Low / flagged
12%
north_eastreview
Snapshot
live
replayreplayable
Decay terraindownload
Higher contours are stronger validity; flatter, lower areas are decaying
High validityAt riskLow / flagged
Validity over time
Share of the base at high validity, last 12 weeks
Decaying premises
Load-bearing items losing confidence
occ.gov MRM94d since validation
48%
Internal note120d
30%
Vendor packunreachable
10%
Events (24h)
4,318
trending_up+6%
By agents
61%
smart_toy2,634 events
PII-touching
12
badgeall warranted
Tamper check
Pass
verified_userappend-only
crisis_alertAnomalies4 signalswatch rules · refer, never edit
PII accessonboarding-svc · agent · 09:20
onboarding-svc accessed PII on KYC-2026-0427 with no declared purpose on file. Bound violated: PII access requires a declared purpose.
After-hours exportA. Reyes (Audit) · person · 02:14
PKG-019 was exported at 02:14, outside the declared working window. Bound violated: exports flag outside 07:00 to 20:00.
Agent authoritycrawler-local · agent · 05:52
crawler-local acted outside its authority bounds: the crawl grant covers intranet /Policies, the run reached /HR before the cut-out. Bound violated: agent actions beyond granted scope.
Volume per principalcopilot-token-04 · agent · today
copilot-token-04 stands at 520 of 500 queries today. Bound violated: daily query volume per principal.
The auditor reads the full trail across every role, the one place where the administrator's operations and the steward's curation are recorded side by side. Reading the trail does not grant access to the knowledge content itself.
Sample size
30
shufflerisk-weighted
Tested
24/30
fact_check80%
Pass rate
88%
check_circle21 of 24
Open findings
3
north_east1 high
Sample & working paperstuneMethod: risk-weightedeventPeriod: Q2 2026
Each item is tested against one criterion; the evidence is the item's own warrant
Findings register
The auditor records a finding and refers it. Remediation is the steward's, never the auditor's.
Packages
12
folder_zipthis year
In preparation
2
edit_notedrafts
Delivered
9
outboxto regulators
Selective disclosure
On
shieldevery export
Assemble a package
PKG-021 · OCC model risk request · draft
Sealing hashes the bundle and signs it with the governance key. A regulator can verify the signature and chain without access to the live base.
Packages
Status and delivery history
Version
v4.2
schedule21d ago
Subscribers
38
trending_up+4 this quarter
Capabilities
11
widgets10 published
Queries (30d)
812k
trending_upvia marketplace
Pack at a glance
What this pack is and how it is built
DomainWealth Management
Co-brandCorvair + Continuum Advisory
Sources1,240 across six trust tiers
Trust mix62% T1 · 24% T2 · 14% T3+
Reference databases8 licensed
How-to guides24
DistributionCorvair marketplace
shield
You run on your own tenancy, with no connection to customer tenancies. You publish to the Corvair marketplace, which distributes releases into each subscribing base and reports subscriber and usage figures back to you.
Recent activity
On the pack, not on any base
Sources
1,240
inventorysix tiers
Tier-1 share
62%
verifiedauthoritative
Pending review
7
inboxin queue
Structure
6
account_treebranches
Sources
The licensed backbone, by trust tier
Review queue
Preview-first: see the cascade, then stage as a batch
Capabilities
11
widgetsshipped
Published
10
check_circlelive
Draft
1
edit_notefor v4.3
Avg. coverage
88%
radaragainst scope
verified_userPortfolio suitability checkPublished
Checks a portfolio against the client's risk profile and the suitability standard, with a warranted rationale.
coverage 94% · 312 sources
badgeKYC enrichmentPublished
Enriches a client record from licensed reference data, redacting PII by policy.
coverage 90% · 188 sources
paymentsFee benchmarkingPublished
Benchmarks fees against the licensed dataset and flags outliers.
coverage 86% · 64 sources
ruleRegulatory suitability mappingPublished
Maps advice to the relevant suitability and conduct requirements.
coverage 88% · 140 sources
insightsClient risk profilingPublished
Builds a risk profile from the questionnaire and holdings, with a cited basis.
coverage 91% · 120 sources
draftCross-border advice narrativeDraft
Drafts a compliant narrative for cross-border advice. In review for v4.3.
coverage 71% · 96 sources
Reference DBs
8
datasetlicensed
How-to guides
24
menu_bookbundled
Licensed sources
340
verifiedunder licence
Last refresh
12d
schedulemonthly
Reference databases
Licensed data that makes the expert authoritative
How-to guides
The procedures that ship with the pack
Current
v4.2
check_circlestable
Next
v4.3
edit_notedraft
On latest
84%
trending_up32 of 38
Changes staged
14
inventoryfor v4.3
Cut a release
v4.3, staged from authoring
Release history
Dated releases and how far they have propagated
Subscribers
38
trending_up+4 QoQ
On latest
32
check_circlev4.2
Behind
6
scheduleon v4.1
Recombining
2
syncapplying v4.2
These figures are reported by the Corvair marketplace. The partner runs on its own tenancy and has no connection to any customer tenancy, so it never sees a subscriber's local content, queries, or audit trail.
Status vocabulary is shared across roles: published when a release reaches the marketplace, available when a subscriber's admin stages it for an environment, applied when their steward recombines it into the base.
Calls (30d)
1.24M
trending_up+11%
Active assets
1,225/1,262
check_circle97%
Hot assets
84
local_fire_departmenttop decile
Unused (90d)
37
ac_unitremoval candidates
Consumption over time
Capability calls per month, marketplace-reported
Active users
Telemetry-derived, live and periodic subscribers
DAU
210
daily active
WAU
740
weekly active
MAU
1,180
monthly active
Stickiness
18%
DAU / MAU
Seats & licensing
Known for every subscriber, including dark sites
Subscribers38
Seats licensed1,420
Seats with telemetry1,260
Queries (30d)812k
Telemetry coverage
How stats reach the partner, across 38 subscribers
Live telemetry
continuous stream
21
Periodic
manual or scheduled uploads
12
Dark (no stats, 12m)
seats licensed, consumption unknown
5
Licensed seats are known for every subscriber, including dark sites, because that comes from the marketplace, not telemetry. Consumption is measured only where telemetry exists, so calls, DAU, WAU, and MAU cover the 21 live and 12 periodic subscribers and exclude the 5 dark sites.
Hot right now
Most-called assets, last 30 days
local_fire_department
Portfolio suitability check
Capability
421k
+18%
local_fire_department
KYC enrichment
Capability
272k
+12%
trending_flat
Fee benchmarking dataset
Reference DB
198k
steady
local_fire_department
Suitability & conduct standard
Standard
156k
+9%
Going cold
Declining or unused, candidates to retire
trending_down
Cross-border advice notes
Method
42k
-19%
ac_unit
Deprecated suitability guide
Guide
12
47d idle
ac_unit
Legacy tax-wrapper note
Source
0
96d idle
ac_unit
Pre-2024 fee schedule
Reference DB
0
140d idle
Settings
Your profile, knowledge bases, notifications, and defaults
Profile & preferences
How you appear and your defaults
Chris Jackson
ckjackson@corvair.ai
Language
Locked to English in this release
EnglishlockOthers
EnglishlockOthers
Interface and answer language are fixed to English for now. Additional languages will unlock these settings.
Appearance
Theme, text size, and answer defaults
CompactDefaultLargeLarger
DetailSummaryBullets
Knowledge bases & drives
What you are connected to
Risk & CompliancePROD
Markets & TreasuryPROD
Notifications & digest
Background jobs and news
Background job completion
Monitored news
Privacy & PII
Governed by your base policy and role · set by your steward and administrator
Default handling for your roleRedact
Detection on ingestionOn · set in the recipe
Purposes you may declareKYC, Onboarding
You cannot change these here. Detection and the redact, block, or allow policy are configured by your steward during base setup and granted per role by your administrator. Where your role permits, you declare a purpose at the point of use; the exemption applies to that one task and is recorded in the warrant.
Access & sign-in
Identity and connected tools
Single sign-oncheck_circleActive
Personal access token