Chatbot (BOT Squad) · Delivery Readiness
Can BOT responsibly commit to its Q3 portfolio — and is there enough sized, staffed, unblocked work to say so?
Capacity itself is not the blocker — with a standard 52 effective-day/quarter assumption, no one is over (busiest: Baghiz Firdausa at 0.66×). The blockers are elsewhere: only 4 of 13 initiatives carry DRI-confirmed man-days (all task-breakdown-derived, none yet a formal sign-off); AI Agent Knowledge (P4) sits at an explicit RFC-review HOLD (4 cross-team sign-off blockers); five initiatives carry a 🔴 blocking external/cross-squad dependency with no committed resolution date from the owning team; and zero initiatives have a hard commitment_date — every date below is the soft target-quarter window only.
commitment_dates in the portfolio — every date is the soft target-quarter windowOne row per initiative: why it can't be locked yet, the single action that closes the gap, and who owns that action. Ordered by the roadmap's ## Priority rank (P1→P14, unranked last). The effort & gate evidence is in §02; the timeline detail is in §03–§04.
| Initiative | Now | Why it can't be committed | What closes it | Owner |
|---|---|---|---|---|
| Autonomous AI Agent — Phase 2 (Refiner) P1 · no date · Phase 1 100% Jira done (80/80), production 2026-06-23 — sandbox-gated, not GA |
Tentative | Fully sized (FE 12 / BE 8 / QA 4.5 = 24.5md, medium confidence) and design-ready (Wulan's live prototype); RFC review PROCEED w/ notes (8.5). Per the task breakdown, chunks 1–8 (Tasks 1.1–2.6, ~22.5 of 24.5md) are executable today against the documented contract with a mocked/stub upstream — Farras, Samudra, Baghiz, Eko, and Tito can start now. Only Task 2.7's real-upstream + config-audit pass is 🔴 blocked, on Data/ML's refine-skill-pack endpoint (no committed date), and per-person split within BE/FE (2 owners each) is not itemized. Phase 1's own hard gate (unresolved autonomous/legacy staging runtime errors since 2026-06-03) blocks Pilot/GA separately and doesn't gate Refiner build. |
Start the BE/FE stub-and-mock slices now (already reflected in Sprint 1); get a date from Data/ML for refine-skill-pack to unblock Task 2.7 only; itemize the BE/FE split between the two named owners each. |
DRI (Dimas) + Data/ML |
| AI Agent Impact Report — Phase 1 P2 · no date |
Do not commit | Now fully sized (FE 15.5 / BE 10 / QA 8 = 33.5md, medium confidence, task-breakdown-derived) and fully staffed (Agus BE, Baghiz FE, Tito/Izzul QA — the squad's standard QA default) — but design is not started (Figma pending), and it depends on Data/BI's per-organization activity-log datamart (🔴 blocking, no date). | Start the Figma design track; get a Data/BI datamart date. Staffing is no longer a gap. | DRI (Dimas) + PM |
| AI Agent Testing — Phase 1 (Historical Validation) P3 · no date |
Tentative | Fully sized (FE 13 / BE 16.5 / QA 8 = 37.5md, medium-high confidence — the best-evidenced initiative in the portfolio) and design-ready; RFC review PROCEED, best score in the portfolio (8.5, zero citation drift after 3 review cycles). Blocked solely on the AI platform squad's batch shadow-inference capacity (🔴, TPM/RPM limits) — no date given. | Get a capacity-allocation date from the AI platform squad for the shared answer-evaluation engine; nothing else is missing. | DRI (Dimas) + AI platform squad |
| AI Agent Knowledge — Phase 1 (Conversation History) P4 · no date |
Do not commit | Sized (FE 1.5 / BE 8 / QA 3.5 = 13md) but its own RFC review verdict is an explicit HOLD (score 7.5) on 4 cross-team/sign-off blockers: webhook contract, masking SLA, Figma design frames, and division-scoping sign-off. Only BE chunks 1–6 + 2b are buildable today; 3 of 7 tasks are blocked. PRD is still `draft`. | Clear the 4 HOLD blockers (REV-2/3/4/5) — two are external Data/AI gates due 2026-07-15, two need internal design/sign-off. | DRI (Dimas) + Data/AI team |
| Unified Agent Scorecard — Phase 1 (Settings & Rubric) P5 · no date |
Park | No effort recorded, no design (concept prompts only, not yet with a designer), no RFC review score yet. DSAI scoring-engine dependency is explicitly non-blocking for Phase 1 per its own PRD — the only bright spot. | Get to a designer and size BE/FE/QA man-days; land an RFC review score. | DRI (Dimas) + PM |
| Unified Agent Scorecard — Phase 2 (Auto-Scoring) P6 · no date · depends on Phase 1 |
Park | No effort, no design, no RFC. Depends on AI platform (DSAI)'s SkillPack evaluator output contract (🔴 blocking, no date) — and structurally can't start ahead of Phase 1. | Sequence behind Phase 1; get a DSAI contract date before sizing. | DRI (Dimas) + AI platform (DSAI) |
| Unified Agent Scorecard — Phase 3 (Analytics Report) P7 · no date · depends on Phase 1/2 |
Park | No effort recorded. Design partially exists (Metabase dashboard reused; host page/nav still needs design). Depends on Data/BI's unified Metabase dashboard + warehouse pipeline (🔴 blocking, no date). | Get a Data/BI pipeline date; finish the host-page design once Phase 1/2 land. | DRI (Dimas) + Data/BI (Analytics) |
| Mekari Action (AI Agent) P8 · backlog-tier · target_quarter 2026-Q3 |
Park | No RFC, no effort, no implementors recorded. `target_quarter` re-baselined to 2026-Q3 (2026-07-09) to match its 26Q3 backlog rank. | Start an RFC; staff it and size man-days. | DRI (Dimas) + TPM |
| Native Integration (AI Agent) — Midtrans Phase 1 P9 · backlog-tier · target_quarter 2026-Q3 |
Park | Has an RFC with a review score (7, PROCEED w/ notes — one mandatory pre-seed fix, two open external confirmations) but no effort man-days and no implementors recorded. Its PRD also carries `jira_epic: BOT-4340` — the one PRD in the repo that would auto-populate per-AC Jira Stories if the (currently untriggered) automated sync ever ran; leave that pipeline untriggered per CLAUDE.md. | Staff it and size man-days; resolve REV-1/REV-7's external confirmations. | DRI (Dimas) + TPM |
| Qontak Action (AI Agent) P10 · backlog-tier · target_quarter 2026-Q3 |
Park | No RFC, no effort, no implementors. PRD header Epic (BOT-3476) drifted from the README's `jira_epic` (BOT-4341) — a wiring gap on top of the sizing gap. | Fix the Epic drift in the PRD header; start an RFC and size the work. | DRI (Dimas) + TPM |
| AI Agent Idle Action (Sequential Idle Follow-up) P11 · backlog-tier |
Park | Best-reviewed unsized initiative in the portfolio — RFC review PROCEED, score 8.5 — but no effort man-days and no implementors recorded yet, purely a staffing/sizing gap, not a readiness gap. | Assign implementors and size man-days — the RFC itself is execution-ready. | DRI (Dimas) + TPM |
| AI Agent Monitoring (Supervisor Alerts) P12 · backlog-tier |
Park | No RFC exists at all yet (PRD only) — the least-progressed initiative in the portfolio besides the off-roadmap MBA initiative. | Start an RFC before any sizing/staffing conversation is possible. | DRI (Dimas) + PM |
| Bot-vs-Human Traffic Split (A/B) P13 · backlog-tier |
Park | RFC exists (draft) with a review file present but no score recorded yet; no effort, no implementors. | Complete the RFC review and size BE/FE/QA man-days. | DRI (Dimas) + TPM |
| Multi-WABA — AI Agent Impact P14 · backlog-tier |
Park | RFC exists (draft) but unreviewed; no effort, no implementors. QA Lane B is driven by an out-of-reach trigger (external Meta WABA platform). | Get the RFC reviewed; size and staff once reviewed. | DRI (Dimas) + TPM |
| Meta Business Agent Integration unranked · off-roadmap gap |
Park | Real 26Q3-targeted initiative (README + 3 phase PRDs exist) but missing from the roadmap Portfolio table (PORTFOLIO_MISSING) and absent from the Priority table — a roadmap-completeness gap, not an active commitment. No `jira_epic`, no RFC, no effort. |
Add a Portfolio row and a Priority rank (or explicitly defer past 26Q3), then start an RFC for Phase 1. | TPM + DRI (Dimas) |
Available-days: 65 eng-days/quarter × 0.8 focus = 52 effective (standard assumption, no per-person override recorded). Autonomous AI Agent Phase 2's BE (Farras+Eko, 8md combined) and FE (Samudra+Baghiz, 12md combined) are not itemized per person in the task-breakdown — split evenly below as a labeled assumption.
| P | Initiative | Target | State | Effort (md) | Gates present |
|---|---|---|---|---|---|
| P1 | Autonomous AI Agent — Phase 2 (Refiner) Farras/Eko BE 8 · Samudra/Baghiz FE 12 · Tito QA 4.5 |
— (2026-Q3 window; Phase 1 100% done, sandbox-gated prod) | Tentative | 24.5 ·med | PRD ✓ RFC ✓ Rev ✓ (8.5) · ~22.5md buildable now ·Data/ML endpoint ✗ date (Task 2.7 only) |
| P2 | AI Agent Impact Report — Phase 1 Agus BE 10 · Baghiz FE 15.5 · Tito/Izzul QA 8 |
— | Do not commit | 33.5 ·med | PRD ✓ RFC ✓ Rev ✓ (8) ·design ✗ · Data/BI datamart ✗ date |
| P3 | AI Agent Testing — Phase 1 (Historical Validation) Farras/Baghiz BE+FE 16.5/13 · Tito/Izzul QA 8 |
— | Tentative | 37.5 ·med-high | PRD ✓ RFC ✓ Rev ✓ (8.5, 0 drift) ·AI platform capacity ✗ date |
| P4 | AI Agent Knowledge — Phase 1 (Conversation History) Eko/Samudra BE+FE 8/1.5 · Tito/Izzul QA 3.5 |
— | Do not commit | 13 ·med | PRD draft · RFC HOLD (7.5) · 4 blockers ✗ |
| P5 | Unified Agent Scorecard — Phase 1 (Settings & Rubric) unstaffed |
— | Park | TBD | PRD ✓ RFC ✓(draft, unreviewed) ·effort ✗ · design ✗ · unstaffed |
| P6 | Unified Agent Scorecard — Phase 2 (Auto-Scoring) unstaffed · depends on Phase 1 |
— | Park | TBD | PRD ✓ RFC ✗ · effort ✗ · design ✗ · unstaffed · DSAI contract ✗ date |
| P7 | Unified Agent Scorecard — Phase 3 (Analytics Report) unstaffed · depends on Phase 1/2 |
— | Park | TBD | PRD ✓ RFC ✗ · effort ✗ · Data/BI pipeline ✗ date |
| P8 | Mekari Action (AI Agent) unstaffed |
— (no date; target_quarter 2026-Q3) | Park | TBD | RFC ✗ · effort ✗ · unstaffed |
| P9 | Native Integration (AI Agent) — Midtrans Phase 1 unstaffed |
— (no date; target_quarter 2026-Q3) | Park | TBD | RFC ✓ Rev ✓(7, notes) ·effort ✗ · unstaffed · REV-1/7 open |
| P10 | Qontak Action (AI Agent) unstaffed |
— (no date; target_quarter 2026-Q3) | Park | TBD | RFC ✗ · effort ✗ · unstaffed · PRD/README Epic drift |
| P11 | AI Agent Idle Action (Sequential Idle Follow-up) unstaffed |
— | Park | TBD | RFC ✓ Rev ✓ (8.5) ·effort ✗ · unstaffed |
| P12 | AI Agent Monitoring (Supervisor Alerts) unstaffed |
— | Park | TBD | RFC ✗ (PRD only) · effort ✗ · unstaffed |
| P13 | Bot-vs-Human Traffic Split (A/B) unstaffed |
— | Park | TBD | RFC ✓(draft) ·Rev score ✗ · effort ✗ · unstaffed |
| P14 | Multi-WABA — AI Agent Impact unstaffed |
— | Park | TBD | RFC ✓(draft) ·Rev score ✗ · effort ✗ · unstaffed |
| — | Meta Business Agent Integration unstaffed · off-roadmap |
— | Park | TBD | RFC ✗ · effort ✗ · no jira_epic · not on Portfolio/Priority tables |
Autonomous AI Agent has the only real, dated milestones (Phase 1 — 100% Jira done, production 2026-06-23, sandbox-gated / not GA) — and, new this pass, its Refiner (Phase 2) BE/FE/QA slices are now buildable with no date needed: per the task breakdown, chunks 1–8 run against a mocked/stub upstream, so Farras/Samudra/Baghiz/Eko/Tito can start today; only Task 2.7's real-upstream pass waits on Data/ML. Every other initiative's forward work is genuinely un-dateable today: three are sized but externally gated with no upstream date (or, for Knowledge, on an explicit HOLD), the rest are unsized/unstaffed. Capacity is not what's withholding a date — Farras/Eko/Samudra/Baghiz/Tito/Agus all have headroom; the remaining gates are external. Hatched grey = exists but undated; solid green = already shipped; solid blue/pink/teal = sized work buildable now (no external gate); dashed red lines mark the two known fixed dates in the whole portfolio (Data/AI's 2026-07-15 webhook + masking-SLA gates for AI Agent Knowledge).
Unlike a data-rich refresh, most rows below still read "gated, no date" — that is the honest state of the portfolio, not a rendering gap. The one exception is Refiner: nearly all of it is buildable now with no date needed. ✓ = shipped/in-flight now · ✓ building = sized work in progress, no external gate · ✗ = blocked, no date · — = not yet staffed.
| Initiative | Backend | Frontend | QA (Lane B) | Lands |
|---|---|---|---|---|
| Autonomous AI Agent — Phase 2 (Refiner) P1 · building now · Lane B |
Farras + Eko 8md combined, unsplit — building now (mocked/stub upstream); only Task 2.7's real-upstream pass is blocked on Data/ML |
Samudra + Baghiz 12md combined, unsplit — building now (mocked BE); same Task 2.7 tail blocked |
Tito 4.5md sized — mocked-BE E2E/regression building now; only the real-upstream staging checklist trails Data/ML |
✓ building, no date needed Phase 1: 100% Jira done (80/80), production 2026-06-23 (sandbox-gated, not GA). Phase 2: ~22.5 of 24.5md executable today against the documented contract; only Task 2.7's real-upstream + config-audit pass is blocked on Data/ML — that portion alone has no date |
| AI Agent Impact Report — Phase 1 P2 · no date · Lane unclassified |
Agus 10md sized, 0.19× load |
Baghiz 15.5md sized, part of Baghiz's 0.66× combined load |
Tito/Izzul 8md sized, part of Tito's 0.46× combined load |
✗ unprovable Effort reconciled (33.5md) and fully staffed; still no design, no date from Data/BI on the datamart dependency |
| AI Agent Testing — Phase 1 (Historical Validation) P3 · no date · Lane B |
Farras 16.5md sized — blocked on AI platform capacity, no date |
Baghiz 13md sized — same block for Task 2.8; most other tasks startable |
Tito / Izzul 8md sized — 11/13 tasks actionable now, 2 gated |
✗ no date Best-specified initiative in the portfolio (RFC review 8.5, zero drift); only missing an AI-platform-owned capacity date |
| AI Agent Knowledge — Phase 1 (Conversation History) P4 · no date · Lane B |
Eko 8md sized — 4/7 tasks actionable now (1.1/2.2/2.4/2.5) |
Samudra 1.5md sized — gated on Figma frames (REV-2) |
Tito / Izzul 3.5md sized |
✗ HOLD RFC review verdict is HOLD (7.5) — 4 blockers, 2 external (Data/AI, due 2026-07-15), 2 internal (design frames, scoping sign-off) |
| Unified Agent Scorecard — Phase 1 (Settings & Rubric) P5 · no date · Lane B |
TBD owner | TBD owner | TBD owner | ✗ not staffed No effort, no implementors, no design yet — concept prompts only |
| Unified Agent Scorecard — Phase 2/3 P6/P7 · no date · depends on Phase 1 |
TBD owner | TBD owner | TBD owner | ✗ not staffed Structurally can't start ahead of Phase 1; DSAI/Data-BI dependencies unresolved |
| Mekari / Native Integration / Qontak Action (backlog tier P8–P10) no date · target_quarter 2026-Q3 |
TBD owner | TBD owner | TBD owner | ✗ not staffed Native Integration has a reviewed RFC (7); the other two have none. All three need a `target_quarter` re-baseline |
| Idle Action / Monitoring / Traffic Split / Multi-WABA (backlog tier P11–P14) no date |
TBD owner | TBD owner | TBD owner | ✗ not staffed Idle Action's RFC is execution-ready (score 8.5) but unstaffed; Monitoring has no RFC at all |
| Meta Business Agent Integration unranked · off-roadmap |
TBD owner | TBD owner | TBD owner | ✗ not on roadmap Not on Portfolio/Priority tables yet; no RFC, no jira_epic |
One change this pass: Autonomous AI Agent Phase 1 is fully closed (100% Jira done, sandbox-gated production) and Refiner (Phase 2) is buildable now with no date needed — Farras, Samudra, Baghiz, Eko, and Tito can start their Refiner slices in July against a mocked/stub upstream. Everything else still reads the same: sized-but-gated or entirely unsized. building = in progress, no external gate · TBD = unsized work that exists but has no forecast.
| PIC · role | July (W1–4) | August (W5–8) | September (W9–13) |
|---|---|---|---|
| Farras BE · autonomous, testing |
Refiner BE — building Autonomous Phase 1 100% done (sandbox-gated prod) frees Farras for Refiner BE now — 8md combined w/ Eko, mocked/stub upstream; only Task 2.7's real-upstream pass trails Data/ML |
Testing BE 16.5md — externally gated on AI platform capacity, no start date |
— |
| Samudra FE · autonomous, knowledge |
Refiner FE — building Autonomous Phase 1 100% done frees Samudra for Refiner FE now — 12md combined w/ Baghiz, mocked BE; only Task 2.7's real-upstream tail trails Data/ML |
Knowledge FE 1.5md — on HOLD (4 blockers, Figma REV-2) |
— |
| Baghiz FE · autonomous, impact-report, testing · 0.66× |
Refiner FE — building 6md half-split w/ Samudra, mocked BE, startable now; only the real-upstream tail (Task 2.7) trails Data/ML |
Impact Report FE 15.5 + Testing FE 13 28.5md known-sized (busiest lane in the squad, still comfortably under 52 effective days) — Impact Report FE (sized, gated Data/BI + design), Testing FE (mostly actionable but gated on capacity date for full E2E) |
— |
| Eko BE · autonomous, knowledge · 0.23× |
Refiner BE — building Action-validation deploy (BOT-4357) now closed with the 100% Jira reconciliation; Eko moves onto Refiner BE — 8md combined w/ Farras, mocked/stub upstream, startable now |
Knowledge BE 8md — 4/7 tasks actionable, but initiative is on HOLD |
— |
| Agus BE · impact-report · 0.19× |
Impact Report BE 10md — sized Blocked on Data/BI datamart — no date; sole assignment this quarter, well inside capacity | ||
| Tito QA · primary · autonomous, knowledge, impact-report, testing · 0.46× |
Phase 1 QA closed + Refiner QA — building Action/Configuration/Tree-Diagram QA shipped 2026-06-17/18; Refiner QA (4.5md) mocked-BE E2E/regression drafting startable now — only the real-upstream staging checklist trails Data/ML |
Knowledge + Impact Report + Testing QA 19.5md known-sized across the three initiatives, all gated on their respective external dependency — comfortably inside 52 effective days |
— |
| Izzul QA · backup · knowledge, impact-report, testing |
No window assignable yet Backup lane behind Tito's primary — nothing to schedule until Knowledge/Impact Report/Testing have dated E2E windows | ||
Cross-cutting constraints — the per-initiative gaps are in the decision table above.
The few portfolio-level moves that unlock the most — in order.