π₯ Inspiration
Sales forecasting is often stuck in the past β static dashboards, manual exports, and endless spreadsheets.
We asked ourselves:
"What if a sales leader could **ask 'what-if' questions directly in Tableau Next* and instantly see the future impact on their pipeline?"*
That idea inspired us to build an Agentic Weighted Pipeline Simulator.
π οΈ How We Built It
- Data Foundation
- We generated a synthetic CRM opportunities dataset (
opportunities.csv) with 3,000 records using Python and Pandas. - Each record includes key CRM fields:
- Region, Segment, Stage, Owner, Amount, Probability, Expected Close Date.
- Region, Segment, Stage, Owner, Amount, Probability, Expected Close Date.
- Formula for Weighted Pipeline:
- We generated a synthetic CRM opportunities dataset (
[ \text{Weighted Pipeline} = \sum_{i=1}^{n} (\text{Opportunity Amount}_i \times \text{Probability}_i) ]
Scenario Layer
- Created a
scenarios.csvwith business-driven adjustments (e.g., increase SMB by +10%, decrease APAC Enterprise by -15%). - Joined with opportunities via a simple
join_keyto dynamically calculate Scenario Adjusted Weighted Pipeline.
- Created a
Visualization Layer
- Built in Tableau Next with:
- KPI Tiles (Total Pipeline, Bookings, Coverage Ratio)
- Scenario-Adjusted Funnel
- Trend Charts (by Expected Close Date)
- KPI Tiles (Total Pipeline, Bookings, Coverage Ratio)
- Enabled scenario filters for region, segment, stage, and win rate adjustments.
- Built in Tableau Next with:
π‘ What We Learned
- Agentic Analytics is more than AI β itβs about making insights conversational and predictive.
- Tableau Next has the potential to go beyond BI dashboards into decision intelligence.
- Synthetic datasets are a powerful way to prototype analytics solutions without access to sensitive CRM data.
π§ Challenges We Faced
- Data realism: Creating synthetic but realistic CRM opportunities required balancing probabilities and stage distributions.
- Scenario logic: Designing the right adjustment rules that are simple yet representative of real sales planning.
- Agentic gap: Tableau Next today is powerful, but bridging to Copilot-like interactions (natural language β filter β recalculation) still requires imagination and future work.
β¨ Impact
This project demonstrates that the next era of analytics is agentic:
- Leaders donβt just want to see what happened β they want to test what could happen.
- By blending Tableau Next + Salesforce Data Cloud + Agentic Simulation, weβve taken a step toward forecasting as a conversation.
Built With
- cloud
- csv
- pandas
- python
- salesforce
- tableau


Log in or sign up for Devpost to join the conversation.