← Back to Case Study
Stage 3 of 7

Planning & PRD

Campaign Operations Optimisation: Phase 1 Salesforce Flows PoC for lifecycle management and SLA automation. Phase 2 positions Agentforce agents as intelligent enhancement.

3.1 Product Requirements Document

PRD: Campaign Operations Optimisation

Campaign Ops Optimisation PRD

Salesforce-native operations layer using Record-Triggered Flows for campaign lifecycle management, complexity scoring, SLA derivation, and auto-assignment. Phase 1 (Complete): Flows-based automation. Phase 2 (Planned): Agentforce agents for intelligent intake validation, routing, and notifications.

Author
Katlego Phokela
Role
Senior Manager: Martech Enablement
Version
V1.0 (Phase 1 Built)
Last Updated
February 2026
Platform
Salesforce Developer Org (PoC)
Status
Phase 1 Complete, Phase 2 Planned

Background & Context

Campaign operations teams manage hundreds of requests per quarter across multiple business units, each carrying different complexity — from reruns with minor changes to net-new multi-channel journeys with audience segmentation and dynamic content. The current process relies on Teams Planner duplication, email threads, and manual estimation. There's no consistent way to measure effort, verify intake quality, or explain SLA breaches with evidence.

This PRD documents a two-phase solution. Phase 1 (completed): Salesforce becomes the single source of truth using Record-Triggered Flows that manage the entire campaign lifecycle — from intake through QA to launch. Flows handle complexity scoring, SLA derivation, auto-assignment, and status management. Phase 2 (planned): Agentforce agents layer intelligent validation, routing recommendations, and proactive notifications on top of the proven Flow foundation.

3.2 Goals & Non-Goals

Goals & Non-Goals

Clear scope ensures the team builds what matters now and deliberately defers what doesn't. Non-goals aren't failures — they're intentional trade-offs showing what we won't build this release.

✅ Phase 1 Goals (Built with Flows)

  • Eliminate manual Planner duplication — all campaign management in Salesforce
  • Complexity scoring: campaign type × channel count × change types → effort score
  • SLA derivation and tracking: auto-calculated from complexity with pause/resume logic
  • Auto-assignment: Flows route to builders with lowest workload, skill-aware routing
  • Standardise QA with structured checklists (replace Excel)
  • Real-time campaign status visibility to BU and Ops
  • Campaign lifecycle management: 7 statuses with validation gates

❌ Phase 2 Goals (Agentforce, Not Yet Built)

  • Agentforce intake validation agent (intelligent completeness checks)
  • Agentforce smart routing agent (AI-powered recommendations with override)
  • Agentforce status notification agent (proactive BU updates)
  • Agentforce QA agent (automated journey validation)
  • Data Cloud integration (currently Flows-only)
  • Marketing Cloud integration (currently Flows-only)
3.3 Feature List (MoSCoW)

Feature Prioritisation

Features ranked using MoSCoW. Phase 1 built with 3 core Flows (Briefing/SLA, Pause/Resume, Initial Assignment Routing) that orchestrate the entire lifecycle. "Must" features are mandatory — without them the product doesn't deliver core value. "Should" and "Could" features enhance Phase 1 and Phase 2 will add Agentforce agents. "Won't" features deferred to Phase 3.

FeatureDescriptionPriorityPhase
Briefing – Calculate Readiness and SLA (Flow)Record-triggered, Before Save. BU fills 3-screen intake form; Intake Validator checks required fields (campaign name, type, channels, audience, BU), blocks incomplete submissions with "Not Ready – BU Action Required"; if valid → Status = Ready for Assignment, Complexity engine calculates Low/Medium/High based on channels (new) or change type (rerun), SLA Target derived with business day logic.Must1
SLA Pause/Resume (Flow)Record-triggered. Pauses SLA clock when Status = "Not Ready – BU Action Required"; tracks paused minutes; resumes and recalculates SLA Target proportionally when work restarts. Ensures fair SLA measurement.Must1
Initial Assignment Routing (Flow)Record-triggered when Status = "Ready for Assignment". Auto-assigns to builder with lowest complexity-weighted in-flight workload. Routes Content Builder for new campaigns, Journey Builder for reruns. Creates Task and sends custom notification.Must1
Campaign Lifecycle Statuses7 statuses: Submitted → Not Ready – BU Action Required → Ready for Assignment → Assigned – In Build → In QA → Ready to Launch → LaunchedMust1
Complexity Scoring FormulaNew campaigns: 1 channel = Low (3d), 2 channels = Medium (5d), 3+ channels = High (7d). Reruns: No changes = Low (2d); Subject/Content/Segment = 1-2 points; Content+Segment/Journey Logic = 3-4 points (Medium/High).Must1
Auto-Created QA ItemsWhen Status = In QA, 10+ standardised Campaign QA Items auto-created: journey validation, data verification, channel testing, consent checks, segment validation, suppression, frequency caps, etc. Each item has Pass/Fail/N/A + evidence field.Must1
Ops Kanban BoardStatus-based list view for standup: filters by Status, shows SLA health (On Track/At Risk/Breached), current assignee, complexity level, days remainingMust1
SLA At Risk & Breached View"Ops – SLA At Risk & Breached" list view with formula field: On Track/At Risk (within 24h)/Breached (past SLA Target). Red highlighting for breached, amber for at-risk.Must1
List Views & KanbanBU views: "My Campaigns". Builder views: "My Content Builds" (new), "My Journey Builds" (reruns). QA view: "In QA (Awaiting Checklist)". Filters by Status, assignee, complexity.Must1
Campaign Intake ObjectSalesforce custom object: Campaign record with fields (Name, Type, Channels, Rerun checkbox, Change Type, BU, Requestor, Status, Complexity, SLA Target, Builder, Launch Date).Must1
Custom Notifications & AlertsBell icon + Outlook/Salesforce notifications to builders on assignment, to requestors on status changes (Not Ready, Ready to Launch, Launched)Should1
Reassignment HandlerTriggered by leave/capacity changes: identifies builder's campaigns, reassigns to next-lowest workload builder, sends notifications via Outlook/SalesforceShould1
SLA Result CaptureAuto-sets "SLA Met" or "SLA Miss" on launch based on actual close date vs SLA Target; requires miss reason if MissShould1
Validation RulesBU required fields enforcement, Not Ready Reason mandatory, Rerun Change Type enforcement, stage transition guardsShould1
Path ComponentVisual tracker: 1–Briefing → 2–Build → 3–QA → 4–Activated; shows current stage and progressShould1
Campaign QA Item ObjectChild object capturing standardised QA checklist; replaces Excel; designed but not yet fully automatedCould2+
Dashboards (3 Audiences)Ops dashboard, BU dashboard, Exec dashboard; designed for Phase 2+Could2+
Demand Planning ViewsForecast capacity based on campaign volume trends; Phase 2+ featureCould2+
Phase 1 Focus: 3 core Flows (Briefing/SLA, Pause/Resume, Initial Assignment Routing) orchestrate the entire lifecycle. 5 mandatory objects/views (Campaign, QA Items, Kanban, At Risk View, List Views) + 5 Should-Have features (Notifications, Reassignment, SLA Result, Validation Rules, Path). Phase 2 will enhance with Agentforce agents and expand Could-Have features.
3.4 The REAL End-to-End Process Flow

How Campaigns Move Through the System (Phase 1: Flows)

New Campaign Flow

1
BU fills Screen Flow form (3 screens)

Screen 1: Basics (campaign name, type, channels). Screen 2: Rerun details (if applicable). Screen 3: Readiness (approvals, content status). Record created with Status = Submitted.

2
Intake Validator Flow fires (record-triggered on creation)

Checks: campaign name, type, channel count, audience criteria, BU name all present. If any missing, Status = "Not Ready – BU Action Required", email sent to requestor with specific missing fields.

3
If validation passes → Status = Ready for Assignment, SLA calculated

Complexity engine runs: new campaigns score by channel count (1=Low/3d, 2=Medium/5d, 3=High/7d); reruns score by change type. SLA Target datetime derived with business day logic. SLA Start timestamp set.

4
Auto Assign Flow fires → finds eligible builder with lowest workload

Complexity-weighted workload calculation: identifies active campaigns assigned to each builder, weights by complexity, finds minimum. Routes Content Builder for new campaigns, Journey Builder for reruns. Creates Task for builder. Custom notification sent.

5
Status = Assigned – In Build

Builder sees campaign in their queue, begins content creation. Campaign record shows builder name, assignment date, SLA status (On Track/At Risk/Breached).

6
When Status moves to In QA → QA Items auto-created

Record-triggered Flow creates 10+ Campaign QA Items with standardised checklist: journey validation, data verification, channel testing, consent checks, etc. Each item has Pass/Fail/N/A status with required evidence field.

7
QA reviewer completes checklist in Salesforce (not Excel)

Reviewer marks each QA Item Pass/Fail/N/A, uploads evidence. Validation rule requires all items completed before Status = Ready to Launch.

8
Ops confirms launch → Status = Launched

Ops lead reviews QA sign-off, confirms activation. Status changes to Launched. SLA result captured: auto-sets "SLA Met" or "SLA Miss" by comparing actual close date to SLA Target. If Miss, capture reason.

Rerun Campaign Flow (Audience + Data Only)

1
BU fills Screen Flow form, marks as Rerun

Same 3-screen form; checks "Rerun" checkbox and specifies change type (No changes / Subject / Content / Segment / Content + Segment / Journey Logic).

2
Intake Validator Flow validates required fields

Checks: campaign name, original campaign reference, change type, audience data file link. Validation rules enforce Change Type field if Rerun=true.

3
Complexity & SLA: Rerun (no changes) = Low, SLA = 2 days

If Change Type = "No changes", complexity = Low, SLA = 2 days. If changes (Subject/Content/Segment/Logic), score by change type: Subject=1, Content=2, Segment=2, Content+Segment=3, Journey Logic=4 points.

4
Auto Assign Flow routes to Journey Builder only

Auto Assign logic skips content builder stage for reruns. Assigns to Journey Builder with lowest workload. Creates Task. Sends notification.

5
Journey builder builds/reactivates, QA, launch

Builder updates journey with new audience data, tests, QA checklist created and completed, launch confirmed. Faster cycle (typically 2 days for no-change reruns).

Key insight: Phase 1 uses Flows for all orchestration — they're deterministic, auditable, and don't require external dependencies. Phase 2 will wrap Agentforce agents around these Flows to add intelligence while keeping Flows as the reliable execution engine.
3.5 User Stories & Acceptance Criteria

User Stories — Real Team Needs (Phase 1)

US-001 · Guided Intake · Must Have
As a BU requestor, I want a guided intake form that captures all requirements upfront, so that I don't get sent back for missing information.
Acceptance Criteria
  • Screen Flow with 3 screens: basics, rerun details, readiness checks
  • Incomplete briefs get Status = "Not Ready – BU Action Required" with email listing missing fields
  • Complete briefs move to "Ready for Assignment" and SLA calculated
  • Validation rules enforce required field submission
US-002 · Auto-Assignment · Must Have
As an Ops lead, I want campaigns auto-assigned to the builder with the lowest workload, so that I don't spend 60 min/day copying to Teams Planner.
Acceptance Criteria
  • Auto Assign Flow triggers on "Ready for Assignment" status
  • Routes based on complexity-weighted in-flight workload calculation
  • Content Builder for new campaigns; Journey Builder for reruns
  • Task created, custom notification sent to assigned builder
US-003 · Content Build Readiness · Must Have
As a content builder, I want to see only campaigns ready for me to build, so that I don't start on incomplete briefs.
Acceptance Criteria
  • "My Content Builds" list view filtered to current user + Status = "Assigned – In Build" + Campaign Type = "New"
  • Campaign record shows all required fields completed (gating prevents display until validation passes)
  • Complexity level and SLA displayed on record for priority
US-004 · SLA Pause/Resume · Must Have
As a journey builder, I want the SLA clock to pause when the BU is causing delays, so that my performance isn't unfairly measured.
Acceptance Criteria
  • When Status = "Not Ready – BU Action Required", Pause SLA Flow triggers, sets Pause Start timestamp
  • When Status moves away from Not Ready, Resume SLA Flow triggers, calculates paused minutes, extends SLA Target proportionally
  • SLA Pause Minutes and SLA Resume Count tracked on record
  • SLA Status formula accounts for paused time in At Risk/Breached calculation
US-005 · Standardised QA · Must Have
As a QA reviewer, I want a standardised checklist in Salesforce so every campaign gets the same quality checks, so that we can stop using Excel.
Acceptance Criteria
  • When Status = In QA, 10+ Campaign QA Items auto-created (journey validation, data verification, channel testing, consent, segment validation, suppression lists, frequency caps, etc.)
  • Each item has Pass/Fail/N/A status with required evidence field (screenshot, test result, approval email)
  • Validation rule requires all items completed before Status can move to Ready to Launch
  • QA records visible in Salesforce for audit and trending
US-006 · At-Risk SLA Visibility · Must Have
As an Ops lead, I want to see at-risk and breached campaigns in one view, so that I can take action before SLAs are missed.
Acceptance Criteria
  • "Ops – SLA At Risk & Breached" list view with SLA Status formula field (On Track/At Risk/Breached)
  • At Risk = SLA Target within 24 hours of Now; Breached = SLA Target passed
  • Shows builder, complexity, current status, days remaining
  • Red highlighting for breached, amber for at-risk
US-007 · EXCO Efficiency Metrics · Could Have
As EXCO, I want to see measurable efficiency gains from this solution, so that I can justify expanding it.
Acceptance Criteria
  • Dashboard showing ops admin time saved (vs. manual Planner process: est. 60 min/day × 250 work days = 250 hours/year)
  • SLA accuracy: % of campaigns meeting SLA Target
  • Workload fairness: variance in effort-weighted assignments across builders (target <15%)
  • Campaign velocity: avg days from Submitted to Launched by complexity level
3.6 Success Metrics

How We'll Know It's Working

Success is measured across input metrics (things we control), health metrics (system stability), and watch metrics (early warnings). North Star: campaign cycle time from brief to go-to-market.

⭐ North Star Metric
Campaign Cycle Time
Target: 30% reduction vs baseline
Business days from brief submission to go-to-market activation. Measures whether standardised process + Flow automation is actually accelerating delivery.
📈 Input Metric
Intake Validation Pass Rate
Target: 90%+ first-time pass
Percentage of briefs passing Intake Validator Flow on first submission without BU needing to resubmit.
📈 Input Metric
Assignment Time
Target: <30 min from validated brief to builder task creation
How fast Auto Assign Flow processes and routes work once brief passes validation.
📈 Input Metric
SLA Pause Accuracy
Target: 100% of paused time correctly calculated and extended
Validates that Pause/Resume Flow logic is working correctly and fairly extends SLA for BU-caused delays.
🟢 Health Metric
SLA Accuracy
Target: 95%+ of SLAs calculated correctly (no false breaches)
Percentage of campaigns with correctly tracked SLAs — complexity scoring working as intended.
🟢 Health Metric
Workload Fairness Index
Target: <15% variance in effort-weighted workload across builders
Shows whether Auto Assign Flow is distributing work fairly. High variance means overload/underutilization.
🟢 Health Metric
QA Consistency Score
Target: 100% campaigns using standardised checklist (not Excel)
Adoption of QA Items in Salesforce. Tracks move away from inconsistent manual QA.
🔴 Watch Metric
Flow Execution Success Rate
Watch: 99%+ success; investigate any failures immediately
Tracks Record-Triggered Flow reliability. Any failures need debugging to ensure automation is trustworthy.
🔴 Watch Metric
Manual Rerun Rate
Target: <2% campaigns requiring manual intervention/rerun of Flows
If high, indicates Flows have edge cases or data issues that need fixing.
3.7 Complexity Scoring Formula

How We Score Campaign Effort

Complexity drives both SLA and workload fairness. The scoring formula is deterministic, transparent, and validated against historical effort data.

New Campaigns

Formula: Channel Count determines complexity level
1 channel = Low (SLA: 3 days)
2 channels = Medium (SLA: 5 days)
3+ channels = High (SLA: 7 days)

Rerun Campaigns

Formula: Change Type determines complexity level
No changes = Low (SLA: 2 days)
Subject line change = Low (1 point)
Content change = Medium (2 points)
Segment change = Medium (2 points)
Content + Segment = High (3 points)
Journey Logic change = High (4 points)

Workload Calculation

For Auto-Assignment: Each builder's in-flight campaigns weighted by complexity
Low = 1 unit
Medium = 2 units
High = 3 units
Total Workload = sum of (campaign complexity × count)
Next assignment goes to builder with lowest Total Workload
3.8 Product Roadmap

Three-Phase Approach: Standardise, Automate, Optimise

Instead of building everything at once, features organised into three phases. Each phase delivers measurable value and provides foundation for the next. Phase 1 standardises process with Flows. Phase 2 adds Agentforce intelligence. Phase 3 scales with prediction.

🟣 Phase 1 — Standardise (COMPLETE)

Briefing – Calculate Readiness and SLA Flow3-screen intake form + Intake Validator; validates required fields; calculates complexity & SLA; triggers on save
SLA Pause/Resume FlowPauses clock on "Not Ready – BU Action Required"; resumes and recalculates target proportionally
Initial Assignment Routing FlowAuto-assigns to lowest workload builder; routes by campaign type; creates Tasks
Campaign lifecycle (7 statuses)Submitted → Not Ready → Ready for Assignment → Assigned – In Build → In QA → Ready to Launch → Launched
Auto-Created QA Items10+ standardised checklist items on In QA transition
Outlook/Salesforce Notifications & AlertsBell + notifications on assignment, status changes
Reassignment HandlerHandles leave/capacity changes with notification
Ops Kanban & SLA At Risk ViewStatus-based list views, SLA health visibility
Complexity Scoring FormulaNew: 1/2/3 channels = Low/Med/High; Reruns: change type-based scoring
SLA Result CaptureAuto-sets Met/Miss on launch; tracks miss reason

🤖 Phase 2 — Automate with Agentforce (PLANNED)

Agentforce Intake AgentIntelligent validation; returns incomplete with specific missing checklist
Agentforce Smart Routing AgentAI recommendations for assignment; Ops lead confirms/overrides
Agentforce Status Notification AgentProactive BU updates at stage transitions with timeline
Agentforce QA AgentAutomated journey configuration validation, test verification
Agentforce Handoff AgentValidates prerequisites between build and journey stages
Dashboards (3 audiences)Ops, BU, Exec dashboards with role-based views

⚡ Phase 3 — Optimise & Predict (FUTURE)

Capacity forecastingPredict workload from submission patterns
SLA tuningData-driven rule adjustments based on 6+ months data
Escalation automationAlert managers when trends shift
Campaign Work Stage (Child Object)Execution tracking separate from intake
Demand planningForecast capacity by BU, by campaign type
Phase philosophy: Phase 1 (Flows) proves the foundation is sound — predictable, auditable, and reliable. Phase 2 layers Agentforce intelligence on top without breaking the core orchestration. Phase 3 scales with prediction and optimization.
3.9 Requirements Summary

Functional & Non-Functional Requirements

Functional Requirements (Phase 1)

IDRequirementPriority
FR-01Briefing Flow (Record-Triggered, Before Save): BU fills 3-screen intake form (basics, rerun details, readiness). Intake Validator validates required fields (campaign name, type, channels, audience, BU). Blocks incomplete with Status = "Not Ready – BU Action Required". If valid, moves to "Ready for Assignment".Must
FR-02Complexity Scoring: New campaigns scored by channel count (1=Low/3d, 2=Medium/5d, 3+=High/7d). Reruns by change type (None=Low/2d; Subject/Content/Segment=1-2pts/Medium; Content+Segment/Logic=3-4pts/High). Stored in Complexity Level field; drives SLA and workload fairness.Must
FR-03SLA Derivation: Derived from Complexity level with business day logic. SLA Target datetime calculated and set. SLA Start timestamp recorded. Rerun (no change) defaults to 2 days.Must
FR-04SLA Pause/Resume Flow (Record-Triggered): Pauses SLA clock when Status = "Not Ready – BU Action Required" (records Pause Start timestamp). When resumed, calculates paused minutes and extends SLA Target proportionally. Tracks SLA Pause Minutes and Resume Count.Must
FR-05Initial Assignment Routing Flow (Record-Triggered): Triggers on Status = "Ready for Assignment". Auto-assigns to builder with lowest complexity-weighted in-flight workload. Routes Content Builder for new campaigns, Journey Builder for reruns. Creates Task and sends custom notification.Must
FR-06Campaign Lifecycle: 7 statuses with validation gates. Submitted → Not Ready – BU Action Required → Ready for Assignment → Assigned – In Build → In QA → Ready to Launch → Launched. Stage transitions enforce validation rules.Must
FR-07Auto-Create QA Items: Triggered when Status = In QA. Creates 10+ Campaign QA Items (journey validation, data verification, channel testing, consent, segment validation, suppression, frequency caps). Each item has Pass/Fail/N/A status + evidence field (screenshot, test, approval).Must
FR-08Ops Kanban Board: Status-based list view for standup. Filters by Status, shows builder, SLA status (On Track/At Risk/Breached), complexity, days remaining. Supports standup workflows.Must
FR-09SLA At Risk & Breached View: "Ops – SLA At Risk & Breached" list view with SLA Status formula (On Track/At Risk within 24h/Breached past target). Red highlighting for breached, amber for at-risk. For proactive SLA management.Must
FR-10List Views & Kanban: BU views ("My Campaigns"), Builder views ("My Content Builds" new / "My Journey Builds" reruns), QA view ("In QA – Awaiting Checklist"). Filters by Status, assignee, complexity for efficient workflows.Must
FR-11Outlook/Salesforce Notifications: Bell icon + Outlook or Salesforce notifications to builders on assignment, to requestors on status changes (Not Ready, Ready to Launch, Launched).Should
FR-12Reassignment Handler: Triggered by leave/capacity changes. Identifies builder's campaigns, reassigns to next-lowest workload builder, sends Outlook/Salesforce notifications.Should
FR-13SLA Result Capture: Auto-sets "SLA Met" or "SLA Miss" on launch based on actual close date vs SLA Target. Requires miss reason if Miss for root cause analysis.Should
FR-14Validation Rules: BU required fields enforcement, Not Ready Reason mandatory, Rerun Change Type enforcement, stage transition guards. Prevents inconsistent states.Should
FR-15Path Component: Visual tracker showing 1–Briefing → 2–Build → 3–QA → 4–Activated with progress indicator. Helps builders see stage in lifecycle.Should

Non-Functional Requirements

IDRequirementTarget
NFR-01PlatformSalesforce Developer Org (PoC); Phase 2+ will add Agentforce, Data Cloud, Marketing Cloud
NFR-02Flow action loggingAll Flow actions (validation, assignment, SLA, notifications) logged with timestamp, actor, justification for audit trail
NFR-03Flow execution timeRecord-triggered Flows complete within 5 seconds; sub-3-second ideal for user-facing actions
NFR-04Concurrent usersSupport 50+ Ops + Builder users + unlimited BU requestors without degradation
NFR-05Mobile-responsiveOps Kanban board, list views responsive on mobile; Screen Flow mobile-friendly
NFR-06Role-based accessBU sees only their campaigns; Builders see assigned queue; Ops sees all; Admin can override all fields
NFR-07Data integrityValidation rules prevent inconsistent state transitions; Process Builder/Flows ensure atomic updates
3.10 Risks & Mitigation

What Could Go Wrong — and How We'll Handle It

RiskLikelihoodImpactMitigation
Flow orchestration — 3 core Record-Triggered Flows managing entire lifecycle Medium Medium Each of the 3 core Flows has a single responsibility: Briefing (validate intake, calculate SLA), Pause/Resume (manage SLA clock fairness), Initial Assignment (route by workload). Well-documented; tested with 5+ test cases per Flow. Deterministic logic reduces risk vs. AI agents.
SLA pause calculation accuracy — Date/Time arithmetic in Flows Low Medium Extensive testing with test data spanning timezones and DST; validation rule checks SLA Target is always in future; manual audit quarterly.
Builder directory maintenance — new joiners/leavers need Campaign Builder records Medium Low Simple admin process documented; Phase 2 will automate with sync from HR system.
Adoption resistance — team used to informal Teams Planner process Medium High Involved team in design; showed time savings in demos; Phase 1 delivers value (visibility, fairness) before Agentforce; Flows are transparent (not a black box).
Data quality in Salesforce briefs — BU submits poor data, validation can't catch it Medium High Intake Validator Flow has required field checks; Phase 2 Agentforce agent will add semantic validation (e.g., audience criteria makes sense); BU training on brief template.
Future Agentforce migration — need to replace Flows with agents eventually Low Medium Clean data model supports either approach (Flows or agents); Phase 2 agents wrap around Flows, not replace them; gradual transition minimizes risk.
Risk philosophy: Phase 1 uses Flows because they're deterministic and auditable. Flows reduce risk vs. Agentforce (no inference, pure logic). Phase 2 Agentforce agents are advisory (Ops can override). If agents fail, Flows provide failsafe execution.
3.11 Stage Summary

PRD Complete: Phase 1 Built, Phase 2 Planned

Key takeaway: Phase 1 (Salesforce Flows) proves the operational foundation is sound: consistent intake, fair complexity scoring, intelligent routing, SLA accuracy, and standardised QA. Flows are deterministic and auditable. Phase 2 (Agentforce agents) wraps intelligence on top without breaking the core system. Phase 3 scales with prediction and optimization.

Phase 1 (COMPLETE) specifies: What Salesforce tracks (campaign brief, complexity, SLA, lifecycle). What Flows orchestrate (validation, routing, SLA pause/resume, QA item creation, notifications). How BU requestors, Ops leads, content builders, journey builders, and QA reviewers interact with the system. What success looks like (cycle time, SLA adherence, workload fairness, Flow reliability).

Phase 2 (PLANNED) will add: Agentforce agents for intake validation (semantic completeness), smart routing (AI recommendations with override), status notifications (proactive BU updates), and QA automation (journey validation). Flows remain the orchestration engine; agents add intelligence.

Next: Stage 4 designs the Salesforce page layouts, Flow diagrams, validation rules, list views, and Phase 1 testing strategy. This PRD is the spec for that detailed design.