How I identified, designed, and built a Salesforce-based solution to fix a broken campaign operations process — Phase 1 PoC complete, Agentforce planned for Phase 2.
A product case study by Katlego Phokela
2026Campaign ops at my organisation ran on a mix of Salesforce, Outlook, Excel, and informal handoffs. There was no consistent way to track campaigns, assign work based on capacity, or know where something was stuck. I mapped the full process, identified where things were breaking, and built a working PoC in Salesforce — 3 custom objects, 3 record-triggered Flows, a complexity scoring formula, auto-assignment logic, and SLA tracking. Phase 1 is live. Phase 2 will explore Agentforce to add intelligence on top of the clean foundation.
| Stage | Focus | Key Deliverable | Status |
|---|---|---|---|
| 1. Discovery & Research | Map the end-to-end campaign lifecycle and pain points | Real process mapping, 5 persona pain points, friction analysis | Complete |
| 2. Strategy & Vision | Define three-phase strategy and product principles | Standardise → Automate → Intelligent three-phase vision | Complete |
| 3. Planning & PRD | Write product requirements and feature specs | PRD with 15 features, complexity formula, SLA rules | Complete |
| 4. Design & UX | Design role-based views and Flow-driven UX | Data model (3 objects), Flow architecture, UI wireframes | Complete |
| 5. Build | Build the working PoC in Salesforce | Working PoC — 3 Flows, validation rules, list views, Kanban, Path | Phase 1 Complete |
| 6. Launch & Rollout | Pilot with one team and plan wider rollout | Pilot plan, training materials, rollout approach | Ready for Pilot |
| 7. Measure & Iterate | Track what's working and what needs fixing | Key metrics defined, feedback approach planned | Planned |
I mapped the full campaign lifecycle — from the moment a business unit submits a brief in Salesforce to when the campaign goes live. I spoke to the ops leads, builders, and BU stakeholders who actually do this work every day to understand where things break down.
The main insight from discovery was that we needed to standardise before doing anything clever. You can't automate a broken process. So I planned it in phases: first get the basics right in Salesforce (Phase 1), then explore Agentforce for smarter automation later (Phase 2).
I wrote the product requirements — what the system needs to do, the rules for complexity scoring, the campaign lifecycle stages, and how assignment and SLA logic should work. I prioritised features using MoSCoW across the phases so Phase 1 focuses on what matters most.
I designed how each role would use the system — ops leads need a pipeline view, builders need their queue, BU requestors need to see where their campaign is. Everything is built with standard Salesforce components: list views, Kanban boards, Path, and record page layouts.
I built the working PoC in Salesforce — 3 custom objects, 3 record-triggered Flows, and the supporting views. The Flows handle: calculating complexity and SLA at briefing, pausing/resuming SLA when waiting on the business unit, and routing new campaigns to the right builder based on workload.
The plan is to pilot with one business unit first — get real feedback, fix what doesn't work, and then roll out to other teams. I've prepared the business case and training approach to support adoption.
The main thing I want to know: are campaigns moving through the pipeline faster and more predictably than before? I've defined the key metrics to track after pilot, and I'll use the data to decide what to improve next.