High-value read-only workflow
Read-only close-exception desk: join synthetic bank transactions, ledger entries, and document indexes; surface unmatched items and an owner digest.
← Back to industry registry
accounting_firms · evidence-backed demand prior · observed 2026-08-27
Accounting Firms
Accounting firms' painful jobs center on document chasing, bank-feed reconciliation and period close; core entities and authority boundaries are known before any client conversation.
45Observed repo slots
8Distinct repos
2Observed demand cells
8Open matrix cells
Demand atoms and workflow
Demand signal / validated demand
E — Catalogue signal: Chase client documents, reconcile bank feeds, and close books on time.
U — Not established in this lane; no attributable industry denominator or repeated accepted-outcome receipt.
Atom join
Specification crosswalk: I — intake_normalize, extract_structure, reconcile_audit, follow_up_chase, report_digest
Industry model
Entities
States
new → normalized → matched | unmatched → needs_review → owner_acknowledged → closed (read model only)
Terminal owner
controller_or_engagement_bookkeeper
Authority boundary
Read-only access to named ledger, bank-feed, and document views; no posting, write-off, reconciliation acceptance, or outbound chase.
Source of truth / exception paths
Source of truth: Accounting/GL plus bank-feed and document-store views; the authoritative system and period owner must be named per pilot.
Exception paths: duplicate document; currency or amount mismatch; stale feed; missing period; rejected permission; conflicting source timestamps
Discovery preparation
Precomputable context
- entity vocabulary and period/close cycle
- likely systems of record (GL, bank feeds, document store)
- archetype fit (case/workflow + reconciliation desk)
- authority boundary defaults (no posting, no write-off)
Questions that must be asked
- which GL/bank-feed/document systems and who owns the period
- close calendar and materiality thresholds
- which exceptions actually cost hours (document chase vs matching vs review)
- client volume denominators (clients, transactions/month)
Contradiction risks
- stated want 'automate the close' vs observed pain usually document collection
- source-of-truth ambiguity between GL and bank feed
Observed repository matrix
These are the non-null rows from repo-matrix-observations.jsonl for this industry. The 1,000-row partition includes reserved cells; only 45 observed repository slots are shown here.
| Repository | Observed slots | Dimensions | Capability families | Observed atoms |
|---|---|---|---|---|
| frappe/books ↗ | 7 | data_model, demand_atom_fit, economics_maintenance, provenance_rights, runtime_deployment, ui_assembly, workflow_behavior | data | extract_structure, follow_up_chase, reconcile_audit |
| akaunting/akaunting ↗ | 6 | data_model, demand_atom_fit, economics_maintenance, provenance_rights, ui_assembly, workflow_behavior | data | extract_structure, follow_up_chase, reconcile_audit |
| darcys22/godbledger ↗ | 6 | data_model, demand_atom_fit, economics_maintenance, provenance_rights, runtime_deployment, workflow_behavior | data | extract_structure, follow_up_chase, reconcile_audit |
| redstreet/beancount_reds_importers ↗ | 6 | data_model, demand_atom_fit, economics_maintenance, provenance_rights, verification_eval, workflow_behavior | data | extract_structure, follow_up_chase, reconcile_audit |
| inforkgodara/store-pos ↗ | 5 | data_model, demand_atom_fit, economics_maintenance, provenance_rights, workflow_behavior | data | extract_structure, follow_up_chase, reconcile_audit |
| juxt/juxt-accounting ↗ | 5 | data_model, demand_atom_fit, economics_maintenance, provenance_rights, workflow_behavior | data | extract_structure, follow_up_chase, reconcile_audit |
| prashants/webzash ↗ | 5 | data_model, demand_atom_fit, economics_maintenance, provenance_rights, workflow_behavior | data | extract_structure, follow_up_chase, reconcile_audit |
| redstreet/fava_investor ↗ | 5 | data_model, demand_atom_fit, economics_maintenance, provenance_rights, workflow_behavior | data | extract_structure, follow_up_chase, reconcile_audit |
Empty-cell gap matrix
The ten category cells retain both demand and supply state. NO_OBSERVED_DEMAND_SIGNAL means the research denominator is empty; it is not a claim that the workflow does not exist.
| Category / empty-cell key | Observed demand | Candidate coverage | Search wave | Candidate examples |
|---|---|---|---|---|
Intake and normalizationintake_and_normalization | NOT_OBSERVED_IN_750_SLOTS0 slots · 0 repos | NO_OBSERVED_DEMAND_SIGNAL10 candidate repos · 10 source rows | NOT_SEARCHEDno targeted search proposal | LimeSurvey/LimeSurvey, ProcessMaker/processmaker, al1abb/invoify, formbricks/formbricks |
Triage and routingtriage_and_routing | NOT_OBSERVED_IN_750_SLOTS0 slots · 0 repos | NO_OBSERVED_DEMAND_SIGNAL31 candidate repos · 31 source rows | NOT_SEARCHEDno targeted search proposal | Alfresco/alfresco-community-repo, Leantime/leantime, ProcessMaker/processmaker, Worklenz/worklenz |
Records, extraction and auditrecords_and_audit | OBSERVED45 slots · 8 repos | SHARED_CANDIDATE_COVERAGE47 candidate repos · 47 source rows | NOT_SEARCHEDno targeted search proposal | Alfresco/alfresco-community-repo, AppFlowy-IO/AppFlowy, BookStackApp/BookStack, BuilderIO/builder |
Classification and prioritizationclassification_and_prioritization | NOT_OBSERVED_IN_750_SLOTS0 slots · 0 repos | NO_OBSERVED_DEMAND_SIGNAL39 candidate repos · 39 source rows | NOT_SEARCHEDno targeted search proposal | Dolibarr/dolibarr, PostHog/posthog, SuiteCRM/SuiteCRM, apache/superset |
Scheduling and coordinationscheduling_and_coordination | NOT_OBSERVED_IN_750_SLOTS0 slots · 0 repos | NO_OBSERVED_DEMAND_SIGNAL14 candidate repos · 14 source rows | NOT_SEARCHEDno targeted search proposal | calcom/cal.diy, alextselegidis/easyappointments, lukevella/rallly, thunderbird/appointment |
Follow-up and communicationfollow_up_and_communication | OBSERVED45 slots · 8 repos | SHARED_CANDIDATE_COVERAGE23 candidate repos · 23 source rows | NOT_SEARCHEDno targeted search proposal | RocketChat/Rocket.Chat, TryGhost/Ghost, activepieces/activepieces, chatwoot/chatwoot |
Reporting and digestreporting_and_digest | NOT_OBSERVED_IN_750_SLOTS0 slots · 0 repos | NO_OBSERVED_DEMAND_SIGNAL15 candidate repos · 15 source rows | NOT_SEARCHEDno targeted search proposal | PostHog/posthog, apache/superset, bigcapitalhq/bigcapital, cube-js/cube |
Monitoring and alertsmonitoring_and_alerts | NOT_OBSERVED_IN_750_SLOTS0 slots · 0 repos | NO_OBSERVED_DEMAND_SIGNAL0 candidate repos · 0 source rows | NOT_SEARCHEDno targeted search proposal | none recorded |
Approval and controlled publishingapproval_and_controlled_publish | NOT_OBSERVED_IN_750_SLOTS0 slots · 0 repos | NO_OBSERVED_DEMAND_SIGNAL21 candidate repos · 21 source rows | NOT_SEARCHEDno targeted search proposal | Budibase/budibase, ProcessMaker/processmaker, activepieces/activepieces, apache/airflow |
Sync and browser operationssync_and_browser_operations | NOT_OBSERVED_IN_750_SLOTS0 slots · 0 repos | NO_OBSERVED_DEMAND_SIGNAL1 candidate repos · 1 source rows | NOT_SEARCHEDno targeted search proposal | activepieces/activepieces |
Shelf candidates joined to this industry
These records come from the permanent source registry's industry-demand join. A joined candidate remains a discovery/research record; qualification and admission are shown explicitly.
| Joined shelf record | Observed slots | Families / atoms | Current verdict | Qualification / admission |
|---|---|---|---|---|
| akaunting/akaunting ↗ | 6 | dataextract_structure, follow_up_chase, reconcile_audit | discovery_candidate | NOT_QUALIFIED / NOT_ADMITTED |
| darcys22/godbledger ↗ | 6 | dataextract_structure, follow_up_chase, reconcile_audit | discovery_candidate | NOT_QUALIFIED / NOT_ADMITTED |
| frappe/books ↗ | 7 | dataextract_structure, follow_up_chase, reconcile_audit | discovery_candidate | NOT_QUALIFIED / NOT_ADMITTED |
| inforkgodara/store-pos ↗ | 5 | dataextract_structure, follow_up_chase, reconcile_audit | discovery_candidate | NOT_QUALIFIED / NOT_ADMITTED |
| juxt/juxt-accounting ↗ | 5 | dataextract_structure, follow_up_chase, reconcile_audit | discovery_candidate | NOT_QUALIFIED / NOT_ADMITTED |
| prashants/webzash ↗ | 5 | dataextract_structure, follow_up_chase, reconcile_audit | discovery_candidate | NOT_QUALIFIED / NOT_ADMITTED |
| redstreet/beancount_reds_importers ↗ | 6 | dataextract_structure, follow_up_chase, reconcile_audit | discovery_candidate | NOT_QUALIFIED / NOT_ADMITTED |
| redstreet/fava_investor ↗ | 5 | dataextract_structure, follow_up_chase, reconcile_audit | discovery_candidate | NOT_QUALIFIED / NOT_ADMITTED |
Open gaps
Intake and normalization
NO_OBSERVED_DEMAND_SIGNAL · 0 observed demand slots · 10 candidate repositories.
NOT_SEARCHED. Exact P03 source-kind coverage for form/portal intake; generic CRM or Stage 0 tags are not counted as intake coverage.
Triage and routing
NO_OBSERVED_DEMAND_SIGNAL · 0 observed demand slots · 31 candidate repositories.
NOT_SEARCHED. Exact P03 workflow/support/project kinds; a product README alone is not semantic coverage.
Classification and prioritization
NO_OBSERVED_DEMAND_SIGNAL · 0 observed demand slots · 39 candidate repositories.
NOT_SEARCHED. Exact P03 CRM/analytics/search candidates are adjacent supply, not proof of a classifier workflow.
Scheduling and coordination
NO_OBSERVED_DEMAND_SIGNAL · 0 observed demand slots · 14 candidate repositories.
NOT_SEARCHED. Exact P03 scheduling kind only; calendar UI alone is not a booking/authority proof.
Reporting and digest
NO_OBSERVED_DEMAND_SIGNAL · 0 observed demand slots · 15 candidate repositories.
NOT_SEARCHED. Exact P03 analytics kind; a reporting surface is not a measured client outcome.
Monitoring and alerts
NO_OBSERVED_DEMAND_SIGNAL · 0 observed demand slots · 0 candidate repositories.
NOT_SEARCHED. No dedicated P03 capability kind is mapped; analytics is not silently promoted to monitoring.
Approval and controlled publishing
NO_OBSERVED_DEMAND_SIGNAL · 0 observed demand slots · 21 candidate repositories.
NOT_SEARCHED. Exact P03 approval/signing kinds; an approval-shaped product does not grant host authority.
Sync and browser operations
NO_OBSERVED_DEMAND_SIGNAL · 0 observed demand slots · 1 candidate repositories.
NOT_SEARCHED. Connector candidates cover a possible sync seam; no direct P03 browser-operation kind is available.
Safe-pilot boundary
Rationale
Useful exception visibility with no ledger mutation; synthetic records can test matching, freshness, tenancy, and evidence without financial side effects.
Falsifier
On a held-out synthetic set, the read model cannot identify mismatches and ownership at a materially better rate than a simple keyed baseline, or no accountable ledger owner can be named.
Sources: research/actionmodel-builder-research-2026-08-26/phase-2/outputs/industry-atom-specifications.md · research/workstreams/p01-client-intelligence/runs/2026-08-27-sprint-1-fable/industry-discovery-priors.jsonl · research/actionmodel-builder-research-2026-08-26/expansion/outputs/repo-matrix-observations.jsonl · research/workstreams/2026-08-28-program-driver/shelf-prebake/industry-gap-matrix.json · knowledge/capability-shelf/source-registry.jsonl