docs: promote AI growth analyst capability

This commit is contained in:
2026-08-21 00:09:21 +03:30
parent 9c88ea603e
commit 59932e1eae

View File

@@ -18,7 +18,7 @@ The hackathon demo will seed one account per merchant. Where merchant profile fi
| 3 | Analytics engine | Deterministic metrics, cohorts, trends, diagnostic drill-downs, bounded parallel on-demand section jobs, and a 30-day Redis cache namespaced by the sole completed immutable dataset snapshot | Data foundation | Yes | | 3 | Analytics engine | Deterministic metrics, cohorts, trends, diagnostic drill-downs, bounded parallel on-demand section jobs, and a 30-day Redis cache namespaced by the sole completed immutable dataset snapshot | Data foundation | Yes |
| 4 | Merchant API | Stable authenticated interfaces and tenant-scoped SSE for partial analytics delivery | Identity and access, analytics engine | Yes | | 4 | Merchant API | Stable authenticated interfaces and tenant-scoped SSE for partial analytics delivery | Identity and access, analytics engine | Yes |
| 5 | Merchant dashboard | Persian RTL dashboard, responsive navigation, accessible charts/tables and explanations | Merchant API | Yes | | 5 | Merchant dashboard | Persian RTL dashboard, responsive navigation, accessible charts/tables and explanations | Merchant API | Yes |
| 6 | AI assistance | AvalAI-backed explanation/recommendation workflow over sanitized aggregates | Analytics engine, merchant API | No | | 6 | AI growth analyst | Deterministic 90-day event-aware forecasts, persisted report snapshots, cost controls, and bounded AvalAI interpretation over sanitized evidence | Analytics engine, merchant API | Yes |
| 7 | Platform operations | Compose, Nginx, manual deployment, phased observability, backup/restore, release and rollback | All deployables | Yes | | 7 | Platform operations | Compose, Nginx, manual deployment, phased observability, backup/restore, release and rollback | All deployables | Yes |
Build sequence: data foundation → identity and analytics → API → dashboard → platform operations → AI assistance. Build sequence: data foundation → identity and analytics → API → dashboard → platform operations → AI assistance.
@@ -45,7 +45,7 @@ Build sequence: data foundation → identity and analytics → API → dashboard
3. Every displayed metric has a definition, source fields, filters, time window, and drill-down path. 3. Every displayed metric has a definition, source fields, filters, time window, and drill-down path.
4. Dataset imports are repeatable and produce a versioned manifest with checksum, counts, timestamps, and validation results. 4. Dataset imports are repeatable and produce a versioned manifest with checksum, counts, timestamps, and validation results.
5. Secrets stay outside Git; examples contain names and safe placeholders only. 5. Secrets stay outside Git; examples contain names and safe placeholders only.
6. AI is post-MVP, receives only minimized/sanitized aggregates, and cannot become the source of record. 6. AI growth analysis is feature-flagged, receives only minimized/sanitized aggregates, and cannot become the source of record. Deterministic forecasts and cached reports remain usable when AvalAI is disabled, unavailable, or over budget.
7. External calls have timeouts and bounded retries; jobs are idempotent and observable. 7. External calls have timeouts and bounded retries; jobs are idempotent and observable.
8. Production changes are reversible and backed by documented health checks and rollback steps. 8. Production changes are reversible and backed by documented health checks and rollback steps.
9. `Paid` and `Verified` are both successful bank payments; `Paid` still requires merchant verification and `Verified` is finalized. Paid-volume and recovery metrics include both while verification metrics keep them separate. 9. `Paid` and `Verified` are both successful bank payments; `Paid` still requires merchant verification and `Verified` is finalized. Paid-volume and recovery metrics include both while verification metrics keep them separate.
@@ -60,6 +60,8 @@ Build sequence: data foundation → identity and analytics → API → dashboard
18. Authentication uses revocable server-side sessions and an opaque host-only `Secure`, `HttpOnly`, `SameSite=Lax` cookie. Passwords use Argon2id; sessions expire after 12 idle hours or seven absolute days and are revoked on logout/account disablement. Demo credentials are generated once and never stored in Git. 18. Authentication uses revocable server-side sessions and an opaque host-only `Secure`, `HttpOnly`, `SameSite=Lax` cookie. Passwords use Argon2id; sessions expire after 12 idle hours or seven absolute days and are revoked on logout/account disablement. Demo credentials are generated once and never stored in Git.
19. Demo authentication stays replaceable and minimal because ZarinPal SSO will own production identity after acceptance. Failed logins are limited by username and IP without permanent lockout; mutations require CSRF and same-origin checks; each user has at most two active sessions. 19. Demo authentication stays replaceable and minimal because ZarinPal SSO will own production identity after acceptance. Failed logins are limited by username and IP without permanent lockout; mutations require CSRF and same-origin checks; each user has at most two active sessions.
20. Demo sessions live in PostgreSQL and store only SHA-256 digests of 32-byte random tokens. Activity writes are coalesced to five-minute windows; Redis remains disposable and is not the session source of truth. 20. Demo sessions live in PostgreSQL and store only SHA-256 digests of 32-byte random tokens. Activity writes are coalesced to five-minute windows; Redis remains disposable and is not the session source of truth.
21. AI growth reports are keyed by authenticated merchant/optional terminal plus data, catalog, analytics, prompt/schema, and model versions. PostgreSQL is authoritative; Redis may coordinate locks, limits, budgets, and hot reads only.
22. Merchant page visits reuse the current matching report. At most one explicit merchant-wide refresh is allowed per Tehran calendar day, and budget exhaustion degrades to deterministic partial output instead of hiding analytics.
## Required capability specs ## Required capability specs