What can you actually do — and where are the gaps?
CAPABILITY MAP: SALES DEV AGENT
════════════════════════════════════════════════════════════
CAPABILITY MATURITY GAP? CATEGORY
────────────────────────────────── ──────── ──── ────────
Email delivery (Resend) █████ (5) Supporting
LinkedIn adapter ████░ (4) Supporting
CRM contacts + deals ████░ (4) Supporting
CRM activities + tasks ███░░ (3) Supporting
Sales Forecasting algo ███░░ (3) Core
Explore-Exploit algo ███░░ (3) Core
SPCL Scoring algo ███░░ (3) Core
Compound Rate Tracking ███░░ (3) Supporting
Document Intelligence ███░░ (3) Supporting
Prospect Research Service ░░░░░ (0) CRITICAL Core
Outreach Composer ░░░░░ (0) CRITICAL Core
Follow-up Sequencer ░░░░░ (0) GAP Core
Lead Scoring Composition ░░░░░ (0) GAP Core
Channel Performance Dashboard ░░░░░ (0) GAP Supporting
Agent Profile (SOUL memory) ░░░░░ (0) GAP Core
Outreach WorkChart ░░░░░ (0) GAP Core
════════════════════════════════════════════════════════════
Maturity Assessment
| Capability | Level | Evidence | Category |
|---|---|---|---|
| Email delivery | 5 — Optimised | Production-ready. Full code path exists. Config only to go live. | Supporting |
| LinkedIn adapter | 4 — Managed | 1043 lines, ~990 real. Mock client only ~50 lines to replace with real API. | Supporting |
| CRM contacts + deals | 4 — Managed | Production, used daily for RFP work. Contact CRUD, deal pipeline. | Supporting |
| CRM activities + tasks | 3 — Defined | Service exists, documented. Needs auto-log wiring for outreach events. | Supporting |
| Sales Forecasting algo | 3 — Defined | Algorithm built and tested. Not wired to any UI or composition layer. | Core |
| Explore-Exploit algo | 3 — Defined | Algorithm built. Needs integration with outreach channel routing. | Core |
| SPCL Scoring algo | 3 — Defined | Algorithm built. Needs integration with outreach quality gate. | Core |
| Compound Rate Tracking | 3 — Defined | Algorithm built. Needs wiring to engagement velocity measurement. | Supporting |
| Document Intelligence | 3 — Defined | Service exists for RFP analysis. Needs composition into prospect research. | Supporting |
| Prospect Research | 0 — Not Built | Nothing exists. Compose from DocumentIntelligence + industry service. | Core |
| Outreach Composer | 0 — Not Built | Nothing exists. Follows RFP answer generation pattern. | Core |
| Follow-up Sequencer | 0 — Not Built | Nothing exists. Uses CRM tasks + email/LinkedIn adapters. | Core |
| Lead Scoring Composition | 0 — Not Built | Wire sales-forecasting + explore-exploit. Composition only. | Core |
| Channel Dashboard | 0 — Not Built | New UI. Data sources exist (webhooks, analytics). | Supporting |
| Agent Profile | 0 — Not Built | Clone from template, populate sales knowledge. | Core |
| Outreach WorkChart | 0 — Not Built | Follows RFP workflow pattern. Orchestration layer. | Core |
Gap Analysis
| Priority | Capability | Why Critical | Investment |
|---|---|---|---|
| P1 | Prospect Research Service | Without this, the agent can't research. No research = no scoring = no outreach. | Build (3 days in S1) |
| P1 | Outreach Composer | Without this, every message is written from scratch. The #1 bottleneck stays. | Build (3 days in S3) |
| P2 | Follow-up Sequencer | Without this, follow-ups still rely on memory. The biggest wait time (5-14d) stays. | Build (3 days in S4) |
| P2 | Lead Scoring Composition | Without this, fit assessment stays as gut feeling. | Wire (2 days in S2) |
| P2 | Agent Profile | Without this, no persistent context across sessions. | Build (part of S1) |
| P2 | Outreach WorkChart | Without this, no end-to-end orchestration. | Build (2 days in S5) |
| P3 | Channel Dashboard | Without this, no feedback loop. But data collection starts with first send. | Build (1 day in S6) |
Investment Strategy
| Category | Count | Strategy |
|---|---|---|
| Core (differentiates) | 10 | Build internally. 6 already at maturity 3, 4 at maturity 0 — these are the sprint. |
| Supporting (enables core) | 6 | Make efficient. 5 at maturity 3-5, 1 at maturity 0 (dashboard). |
| Generic (commodity) | 0 | Nothing to buy. Platform already built. |
The 80/20 split: 80% of components exist at maturity 3+. The sprint is wiring (composition) + building the 20% that's new. This is the "9 algorithms, 0 demand" problem — the parts exist, the machine doesn't.
Dependency cross-check: Both P1 gaps (Prospect Research, Outreach Composer) have only soft dependencies — no hard blockers from the Dependency Map. The hard dependencies (LinkedIn app approval, Identity PRD) affect supporting capabilities, not the critical path to first value. Engineering can start on P1 gaps immediately.
Gate
Before executing:
- All capabilities assessed with evidence-based maturity — YES (16 capabilities, all with evidence)
- Capabilities categorised (Core / Supporting / Generic) — YES (10 Core, 6 Supporting, 0 Generic)
- Gaps identified and prioritised (P1 / P2 / P3) — YES (2 P1, 4 P2, 1 P3)
- Investment strategy matches category — YES (build core, make supporting efficient)
- Action plan has owners — YES (all mapped to sprint tasks)
- Critical gaps have escalation paths — YES (P1 gaps are first two sprints)
Context
- Dependency Map — Previous: what must happen first
- Agent & Instrument Diagram — Next: how do agents orchestrate
- Capability Map Template — The empty pattern