Dashboard

The Analytics entry in the sidebar (formerly "Dashboard") is your read-only view of how Sales is doing — what's being captured, what's stuck, what's converting, and what AI is costing.

Unlike Todo, which is for daily execution, Analytics is for stepping back. It's designed for:

• Reps reviewing their own queue, response times, and outcomes
• Managers spotting where coaching matters and where the pipeline is stuck
• Founders tracking revenue, AI spend, and forecast posture

How the page is laid out

Analytics is a single page with three regions:

1. Sticky filter bar at the very top — date range, scope (Me / My Team), optional rep filter for managers.
2. Side-nav rail on the left listing the sections — clicking jumps to that section, scrolling updates the active highlight automatically.
3. Section panels stacked in the main area. Each panel can be collapsed independently; the open/closed state is preserved in the URL so reloading or sharing a link keeps your view.

On screens narrower than 1024px the side-nav collapses into a horizontal chip strip at the top of the content area.

Sharable URLs

The page state lives in the URL (?range=30&scope=team&rep=…§ion=pipeline&expanded=execution,outcomes). Copying the URL and sending it to a teammate lands them on the same view, with the same filters, scrolled to the same section. Bookmarks behave the same way.

Filters

What's in each section

The section order is role-aware: reps see their own queue first, managers see pipeline + team performance first.

Today (always at the top)

A sticky strip of seven KPI tiles answering "what needs attention right now":

• Customers Waiting — leads with an unanswered inbound message
• SLA Breaches — leads past the workspace's first-response window
• High-Priority Pending — top-of-queue todos not yet handled
• Wins This Period + Revenue Won — both with a vs-previous-period delta
• Win Rate / Avg Cycle — % won and average time-to-win
• Open Pipeline — count of active leads + total open value

Pipeline Reality

• Operating Funnel — Signals → New Leads → Waiting Now → Pending Todos → Wins.
• Stage Health — counts, value, and median age per active stage. Click a stage row to open Pipeline filtered to that stage.
• Stage Decisions + Top Transitions — what kinds of stage moves happened in the window, and the most common from→to transitions.
• Stage Velocity (median + p90) — how long deals actually spend in each stage. Computed from completed transitions over a fixed 180-day window so the p90 stays stable regardless of the date range filter. Stages with under 3 samples show a "not enough yet" hint instead of a misleading number.
• Stale Lead Alert — top 10 active leads whose age in their current stage exceeds that stage's p90. Click a row to open the lead.

Execution & SLA

• Todo Velocity — created vs completed todos per day, plus pending-now and median response time.
• Priority Pressure — pending high / medium / low and a callout for leads outside the SLA window.
• Handling Mix — completed vs snoozed vs dismissed for the period, plus average response time.
• First-Response SLA Distribution — how response times bucket out: under 15 minutes, 15-60 minutes, 1-4 hours, 4-24 hours, over 24 hours. Sample size shown so you know how trustworthy the shape is.
• Engagement Decay — active leads bucketed by days since last interaction (<3d / 3-7d / 7-14d / >14d). Distinct from Stale Lead Alert: that one tracks stage-stuck time, this one tracks touch frequency. Active leads with no recorded touch are counted separately as a footnote so they don't get conflated with leads touched today.

Coverage & Trust

• Signal Volume + Signal Sources — total signals captured, broken down by source.
• Inbound Signal Mix by Source (stacked daily) — top 4 sources plotted individually with everything else in "other". Useful for spotting channel shifts.
• Trust Checks — stage coverage %, decision coverage %, inbound waiting without a todo, pending stage reviews.
• Won-Deal Signal Pattern — for the cohort of deals that closed-won in the last 90 days, the signal families that touched them in the 30 days before close. The closest thing to a data-driven "what works" without explicit experiments.

Team Performance (manager-only)

• Todos per Rep — distribution of new todos in the period across the team.
• First-Response SLA per Rep — median response time per rep, sorted slowest first so coaching attention lands where it matters.
• Draft Acceptance per Rep — for each rep in the period, the percentage of AI drafts they sent unchanged (or with no separate final_message recorded). Sorted highest first. The workspace-wide value also lives on each rep's coaching card below as a single number; this widget is the period-scoped, side-by-side comparison.
• Rep Snapshots — per-rep coaching cards with win rate, pending todos, response time, draft acceptance rate, and an AI-generated summary. Clicking opens Pipeline filtered to that rep.

Outcomes, AI Cost & Forecast

• Outcome Changes — won / lost daily bars + revenue won.
• Current Lead State — distribution across need-attention / in-progress / won / not-responding / lost.
• Lost Reasons — why deals lost in the window.
• AI Cache Hit Rate — what % of AI runs reused a cached result instead of making a fresh model call. Includes an estimated USD saved (avg vendor cost × cache hits).
• AI Consultations per Rep — how often each rep asked the AI for help, with the cost.
• AI Activity & Spend — runs, internal spend, daily cost, and a breakdown of what AI work types ran.
• Forecast — only for workspaces on the Revenue plan. Weighted forecast value, commit / stretch bands, recent close pattern, weighted-by-stage and per-rep outlooks. Workspaces without the entitlement see an upgrade card here.

Drill-downs

Almost everything in Analytics is clickable. The destinations:

Middle-click and Cmd/Ctrl-click open the destination in a new tab.

Realtime updates

Analytics subscribes to leads, todo_actions, lead_stage_decisions, pipeline_stage_suggestions, and pipeline_stages realtime channels. When something changes that affects a section's data, the section silently refetches in place — no full-page reload, no flicker. There's also a 2-minute background refresh as a safety net.

Key terms

Tips

For reps

1. Open Today first thing in the morning to see your queue, SLA breaches, and what closed yesterday.
2. Use Execution & SLA to see where your response times sit in the distribution. If you're mostly in the >24h bucket, that's worth a conversation.
3. Stale Lead Alert in Pipeline Reality surfaces specific leads that have been stuck longer than your stages typically run — work those before they go cold.

For managers

1. Pipeline Reality is your morning check — Stage Velocity tells you which stages are slowing down before the funnel does.
2. Team Performance's "First-Response SLA per Rep" is sorted slowest first on purpose. That's where coaching is overdue.
3. Won-Deal Signal Pattern is the closest thing to a data-driven "what works" — if meeting is consistently in the top 3 for won deals, that's an argument for more meetings.

For founders / executives

1. Today's KPI strip is your daily glance — Wins, Revenue Won, Win Rate.
2. Outcomes section's AI Cache Hit Rate is the AI efficiency number; if it's trending down, the runner is making more fresh calls than necessary.
3. Forecast (when entitled) gives you Commit / Stretch bands plus the assumptions used to compute them — read those when the numbers feel off.

What not to do here

Analytics is read-only by design:

• Don't look for individual lead actions here — use Todo.
• Don't try to edit pipeline state here — use Pipeline.
• Don't dig into raw AI traces here — use Activity (drill-downs from Analytics charts will route you there with filters pre-applied).