SalesPulse Live

A Bloomberg-terminal-style B2B SaaS sales pipeline command center delivered as a static GitHub Pages site. It's the canvas for a live AI-assisted agent-vs-agent demo loop: a Critique Agent files feedback as GitHub Issues, a Builder Agent implements them, the site auto-redeploys, the audience watches it evolve in real time.

Synthetic data only. No backend, no database, no auth, no build step.

---

🔗 Live links

Thing	URL
Live dashboard	https://hashwanthven.github.io/salespulse-live/
Feedback form (share with audience)	https://hashwanthven.github.io/salespulse-live/requests.html
GitHub repo	https://github.com/HashwanthVen/salespulse-live
Issues (where critique lands)	https://github.com/HashwanthVen/salespulse-live/issues
Actions / deployments	https://github.com/HashwanthVen/salespulse-live/actions
Dashboard context (source of truth for agents)	.github/agents/dashboard-context.md
Critique Agent guide	.github/agents/critique-agent.md
Builder Agent guide	.github/agents/builder-agent.md

Cache-busting during demos: append ?v=demo-N (e.g. https://hashwanthven.github.io/salespulse-live/?v=demo-2).

---

🎯 What this dashboard is

A polished, dense, terminal-style view of one quarter of synthetic sales pipeline data, including:

• 6 headline KPIs — Pipeline Value, Weighted Pipeline, Win Rate (TTM), Avg Deal Size, Sales Cycle, Pipeline Coverage
• Pipeline Funnel — 6 stages with stage-to-stage conversion rates
• Forecast vs Quota — weekly cumulative commit and best-case lines vs a quota target
• Top Open Deals — 12 opportunities, filterable by region and forecast category
• Rep Leaderboard — quota attainment, pipeline, closed-won count, activity score
• Pipeline by Segment (Enterprise / Mid-Market / SMB) and by Region (NA / EMEA / APAC / LATAM / MEA)
• Deal Risk & Anomaly — stalled deals, single-threaded deals, pushed close dates
• AI Insights — rotating executive narrative
• Live Audience Feedback — reads GitHub Issues directly via the public REST API, auto-refresh every 30 seconds
• A scrolling ticker, function-key nav, command line (HELP, GO DEALS, FILTER NA, FCST BEST), and status bar

The aesthetic is intentional: green-on-black, monospace, sharp borders, dense panels, tabular numerics. Mirrors what a real RevOps / CRO floor would project on a wall.

---

👥 Who it's for

Persona	What they want to see in 5 seconds
VP Sales / CRO	"Are we going to hit number?" — commit vs quota gap, coverage, who's behind
RevOps / Sales Ops	Funnel conversion, stalled deals, forecast hygiene, rep activity outliers
Sales Manager	Their team's leaderboard slice, their top deals, deal-level risks
Account Executive (AE)	Their own deals and where they stand on attainment
CFO / Finance	Forecast accuracy and confidence around the commit number
Marketing leader	Top-of-funnel volume and qualification rate
Board member	A glance: are we ahead/behind, by how much, and what's the trend

Primary persona for v0.1: VP Sales / CRO. Secondary: RevOps.

---

🤖 The agent-vs-agent loop

┌────────────────────┐     files GH issues     ┌────────────────────────────┐
│  Critique Agent    │ ───────────────────────▶│  HashwanthVen/             │
│  (reads live URL,  │                         │  salespulse-live/issues    │
│   §9 heuristics,   │                         └────────────────────────────┘
│   max 3/pass)      │                                       │
└────────────────────┘                                       │ picks oldest
        ▲                                                    ▼ high-priority
        │ next critique pass                       ┌────────────────────┐
        │ reads new state                          │  Builder Agent     │
        │                                          │  (implements,      │
        │                                          │   commits, push)   │
        │                                          └────────────────────┘
        │                                                    │
        │            ┌───────────────────────────────────────┘
        │            ▼
        │    push to main
        │    .github/workflows/pages.yml deploys
        │    https://hashwanthven.github.io/salespulse-live/ updates
        │
        └─── loop forever (or until issue backlog ≥ 8) ───┘

Both agents read the same source-of-truth before acting: .github/agents/dashboard-context.md. That file pins down:

• KPI definitions (so changes don't drift)
• Forecast categories (commit / bestcase / upside / omitted)
• Funnel stages (prospects → qualified → discovery → proposal → negotiation → closed-won)
• Persona priorities
• "What good looks like" / "what's not good"
• 8 critique heuristics, ranked

---

🚀 Run locally

No build step. Just open the file:

start index.html

Or serve it:

python -m http.server 8080
# visit http://localhost:8080

---

☁️ Deploy to GitHub Pages

.github/workflows/pages.yml deploys the repo root to GitHub Pages on every push to main and on manual dispatch. Pages source is already configured to "GitHub Actions" for this repo.

To recreate the setup elsewhere:

1. Repository → Settings → Pages
2. Source → GitHub Actions
3. Push to main. Done.

---

📝 Submitting feedback

The audience-facing form at requests.html is intentionally tiny — 3 fields, mobile-first:

1. What should we change or add? (title — required)
2. Tell us more (description — optional)
3. Type (Feature / Bug / UI polish / Insight)

Tap ▶ SUBMIT → new tab opens at GitHub with everything prefilled → tap "Submit new issue" → it lands at https://github.com/HashwanthVen/salespulse-live/issues.

Requires a free GitHub account. Once landed, the issue is fair game for the Builder Agent to implement.

---

🎬 Suggested live demo flow

1. Open the live URL — let the ticker scroll and panels populate.
2. Walk the audience through the persona of the CRO at a Friday forecast call.
3. Open the Feedback form, capture an audience suggestion (or point them at the URL on their phones).
4. Watch the issue appear at /issues and in the Live Audience Feedback panel on the dashboard within 30 seconds.
5. Trigger the Builder Agent on that issue (point it at .github/agents/builder-agent.md).
6. Watch the commit land, the workflow deploy, and the dashboard update live.
7. Trigger the Critique Agent (point it at .github/agents/critique-agent.md) to file the next round of feedback.

---

📂 Repo layout

File	Purpose
index.html	Main dashboard
requests.html	Mobile-first 3-field feedback form
styles.css	Green-on-black terminal theme
app.js	Dashboard behavior — ticker, clock, KPIs, funnel, chart, deals, reps, command bar, audience feed
requests.js	Form submit + live feedback feed
data.js	All synthetic sales data
.nojekyll	GitHub Pages serves files as-is
.github/workflows/pages.yml	Auto-deploy on push to main
.github/ISSUE_TEMPLATE/feature_request.yml	Structured feature form
.github/ISSUE_TEMPLATE/bug_report.yml	Structured bug form
.github/agents/dashboard-context.md	Source of truth for both agents (read this first)
.github/agents/critique-agent.md	Critique Agent operating guide
.github/agents/builder-agent.md	Builder Agent operating guide

---

Synthetic data only. Built for live AI-assisted development demos.