Agentic KPI Concierge

An AI-powered Tableau Next dashboard that unifies revenue, churn, support, and customer satisfaction insights into one interactive view.

Comment

Inspiration

SaaS companies constantly face a triple challenge:

  • Sustaining revenue growth
  • Minimizing customer churn
  • Maintaining customer experience (CX)
    We wanted to build a tool that goes beyond static reporting by combining a semantic layer, natural-language Q&A, and business storytelling. Our goal: Make executives feel like they have a concierge agent for KPIs, always ready to answer questions and highlight risks.

What it does

The Agentic KPI Concierge Dashboard unifies Revenue, Churn, Support, and Targets into one semantic-driven model:

  • Provides executive summary KPIs (Total Revenue YTD, Churn Risk, Avg CSAT, Ticket Volume).
  • Exposes Ask a Question (Q&A) with suggested prompts like “Churn risk by product” or “Revenue gap vs target.”
  • Visualizes trends and gaps (Revenue by region/product, Churn by plan, Support vs CSAT).
  • Embeds narrative insights + business impact so leaders can move from data → decisions quickly.

How we built it

  1. Data Modeling
    • Cleaned and joined datasets for revenue, churn, CSAT, and ticket volume.
    • Built a semantic layer with key dimensions (Region, Product, Month) and metrics (MRR, Targets, Churn Risk, CSAT).
  2. Dashboard Development
    • Used Tableau Next to create KPI cards, visualizations, and narrative containers.
    • Added Ask Data / Q&A panel with domain-specific starter questions.
  3. Business Storytelling
    • Linked each visual to problem-insight-impact, making the dashboard both analytical and persuasive.

Challenges we ran into

  • Filters: Adding Region/Product/Month filters globally wasn’t straightforward due to Q&A/semantic constraints.
  • Ask Data Panel: Although enabled, some Q&A features were grayed out or inaccessible.
  • Time constraints: Couldn’t add predictive ML churn modeling, which would have boosted foresight.
  • Design balance: Keeping dashboards clean but still information-rich required iteration.

Accomplishments that we're proud of

  • Integrated Q&A natural language directly in dashboards with “starter prompts.”
  • Created narrative business insights inside Tableau, rarely done in typical BI projects.
  • Identified a clear $3.2M annual uplift if churn in Standard plans is reduced by just 10%.
  • Designed a multi-page executive → detail drilldown experience: Executive Summary, Revenue Insights, Churn Insights, Support, and Targets vs Actuals.

What we learned

  • How to blend semantic modeling + visualization + business narrative into one solution.
  • That churn and support tickets are tightly correlated with CSAT, and framing that insight in dollars resonates with decision-makers.
  • Hackathon dashboards need both analytical accuracy and storytelling creativity to stand out.

What's next for Agentic KPI Concierge

  • Add interactive filters (Region, Product, Month) across all visuals.
  • Integrate predictive churn modeling (ML-based) to anticipate retention risk.
  • Expand Q&A library with deeper semantic definitions for executives.
  • Enable real-time data refresh and exportable narrative reports.
  • Explore generative AI summarization of KPI insights.

Team

Solo: Sweety Seelam

Built With

  • cloud-storage
  • saas
  • semantic-layer-modeling
  • sql
  • tableau-next
Share this project:

Updates