With the problems mapped, I needed to define what we’re building with Agentforce, why it matters, and how it transforms campaign operations from manual coordination to intelligent automation.
I needed a clear statement to guide every decision — if a feature didn’t serve this vision, it didn’t belong in the product. The vision had to include Agentforce as the intelligence layer that transforms manual operations into proactive, coordinated workflows.
| Element | What I Decided | Why |
|---|---|---|
| Who it’s for | Ops leads, content builders, journey builders, QA reviewers, BU marketing managers, EXCO | Each role has different daily questions and pain points — the product needs to serve all of them |
| What they need | End-to-end campaign visibility from Salesforce brief through Data Cloud segmentation to Marketing Cloud activation — without manual tool-switching or status chasing | Today the team context-switches between tools and manually coordinates handoffs. This visibility gap creates delays and rework. |
| What the product is | A two-phase solution: (1) Salesforce Flows PoC delivering standardised intake, auto-assignment, and SLA governance today, (2) Agentforce AI agents for intelligent routing, validation, and notifications in Phase 2 | Phase 1 is working today with Flows. Phase 2 is the planned evolution to Agentforce intelligence. This combination gives us governance + intelligence + execution in one integrated stack, all Salesforce-native |
| Key differentiator | Working PoC that already eliminates Teams Planner and delivers fair, data-driven assignment — with a clear path to Agentforce intelligence | Not a presentation or concept. A working solution deployed today with validated business rules. Phase 2 agents will enhance, not replace |
| Instead of | Manual Teams Planner copies, email-based handoffs, Excel QA checklists, informal chat-based coordination | These are what the team uses today, and they don’t scale or leave audit trails |
| Core principle | Standardise first, then automate, then make intelligent — each phase delivers value independently | Too many AI projects fail because they try to automate broken processes. We fix the process first, layer in agents second. |
I mapped out what each persona is trying to do, what frustrates them, and how the Agentforce-powered product addresses it — to make sure I wasn’t building features nobody asked for.
These are the rules I used to make decisions throughout the project. Whenever I wasn’t sure whether to include a feature or how to design something, I came back to these.
Fix the process before automating it. Clean workflows make AI effective. This is Phase 1.
Flows handle routine routing and validation. Ops leads keep judgment for exceptions. Phase 2 agents will enhance this, not replace it.
Salesforce → Data Cloud → Marketing Cloud as one pipeline. No more tool switching or manual handoffs.
Derived complexity scoring sets SLAs, assignment routing, and capacity planning. Trustworthy because it’s calculated, not guessed.
Agentforce agents flag issues before they become escalations. Status updates go out automatically, not manually.
Every agent action, assignment decision, and stage transition is logged and traceable. Compliance by default.
The product is designed to deliver value incrementally. Each phase builds on the previous one and adds a new layer of operational capability. The key: standardise the foundation first, then layer in Agentforce intelligence.
| Phase | Scope | Why It’s Safe |
|---|---|---|
| Phase 1: Standardise (Foundation) | COMPLETE. Structured intake via Screen Flow (replaces Teams Planner), campaign lifecycle, complexity scoring, auto-assignment via lowest-workload routing, SLA governance, standardised QA checklist, Kanban board, list views | No AI yet — just fixing the process and data foundation. Already built. Immediate value. Ops stays in full control. |
| Phase 2: Automate with Agentforce (Intelligence) — PLANNED | Agentforce Intake Validation Agent (checks requirements completeness), Agentforce Smart Routing Agent (assigns based on complexity + workload + skills), Agentforce Status Notification Agent (proactive BU updates), Agentforce Handoff Agent (validates prerequisites between stages) | Agents recommend and assist — Ops can override. Each agent solves one specific, well-defined problem. |
| Phase 3: Optimise & Scale (Predictive) — FUTURE | Agentforce QA Agent (automated journey validation replacing Excel buddy system), demand forecasting (planned vs actual campaign volume), capacity planning with AI-recommended team scaling, Agentforce Campaign Insights Agent (pattern recognition across campaign performance) | Built on clean data from Phase 1 and proven agents from Phase 2. Solves the demand vs capacity challenge. |
The team has seen project management tools before. Here’s why this approach is different.
Not a presentation or concept. A working Salesforce Flows PoC with 3 custom objects, 3 Flows, and validated business rules.
Most teams try to automate broken processes. We fix the process first, then layer intelligence on clean data. This is why Phase 1 matters.
Salesforce PoC today, Data Cloud + Marketing Cloud integration in future. No third-party glue, no data silos, no manual handoffs between tools.
Every downstream decision (SLA, routing, capacity) derives from structured complexity scoring. This makes the operation fair, predictable, and auditable.
Built to demonstrate AI ROI: time saved by Ops, SLA accuracy, assignment fairness, campaign cycle time reduction, team confidence in the operation.
| Question | My Answer |
|---|---|
| What’s the vision? | Standardised Salesforce campaign operations with Flows delivering immediate value, evolving to Agentforce intelligence |
| Who am I building for? | Ops team, content/journey builders, QA reviewers, BU stakeholders, EXCO |
| What makes it different? | Working PoC with Flows delivering standardised intake and auto-assignment today. Phase 2 planned for Agentforce agents |
| What platform? | Salesforce MCP (Phase 1 complete), Data Cloud + Marketing Cloud integration planned |
| How do I measure success? | Ops admin time reduction, SLA accuracy, assignment fairness, campaign cycle time, EXCO confidence score |
| Delivery approach? | Three phases: Standardise (Phase 1) → Automate with Agentforce (Phase 2) → Optimise & Scale (Phase 3) |
| Core principle? | Standardise first, then automate, then make intelligent. Each phase delivers value independently. |
| What comes next? | Phase 1 COMPLETE (Salesforce Flows), Phase 2 PLANNED (Agentforce agents), Phase 3 FUTURE (predictive scaling) |