AI moat · Ops

The dashboard that
tells you what to do next.

A live event dashboard for the organiser, refreshing every 30 seconds. Check-in throughput and ETA to clear the queue. Sessions at 80% and 100% capacity. Sponsor booths hot, warm, cold, silent. And on top of the numbers, an AI operator that reads the situation and hands you one concrete next action — open a third check-in counter, or ask the MC to plug Booth 14 — in two sentences.

Live · Ops
TechSummit '26
Refreshed 10:47:22 · every 30s
🧠 AI OPERATOR
Check-in is running at 12/min and should clear in 8 minutes. Move a volunteer from Gate 2 to Gate 3. Booth 14 still at zero — ask the MC to plug it before session 2.
Check-in
412 in 96 left 12/min
Sessions
Keynote FULL · AI Track 92%

Everything the organiser stares at on event day.

🧠
AI narrative
Two sentences. What's happening + what to do about it. Regenerated every 60 seconds via your Claude/OpenAI/Gemini/Groq chain. Cached to keep API cost near zero.
📈
Check-in pulse
Checked-in, remaining, per-minute rate over the last 15 min, projected minutes to zero. Status pill turns amber when throughput slows below expected, red when idle.
🎤
Session capacity
Every session listed with RSVP-vs-capacity bar. High (80%+) and Full (100%+) get colour-coded so you can flip rooms early instead of at the door.
🏪
Sponsor velocity
Top 5 booths ranked by 30-min scans. Silent booths list separately with a nudge to have the MC name-drop them. Sponsor renewal starts here.

Because event day is a WhatsApp storm you can't win.

Ask any conference organiser what event day feels like. It's a WhatsApp storm. "Gate 2 has a queue." "Keynote room is full." "Booth 14 says no one's coming by." "Where's the second scanner?" Every message wants a decision. You're standing in the hallway triaging by memory.

The Onsite AI Operator turns the storm into a single tab. Live queue rate is on screen — you don't need someone to WhatsApp you "it's slow." The full session appears in red 15 minutes before the door — you flip rooms without a shouting match. The silent booth shows up on its own — the MC gets a note between sessions.

The AI narrative is the difference. It doesn't just show numbers; it reads them the way a seasoned event runner would and tells you the one thing to do next. Two sentences, updated every 30 seconds. That's the copilot every organiser has wanted since we started building this.

Where live ops matters most.

📷
Every check-in feeds the throughput chart. Combined with offline scanning, the number is right even when a scanner briefly loses signal.
🎟
Capacity alerts come from the same RSVP data that powers the public agenda page. One source, no double-entry.
💼
Scan velocity data on the live dashboard becomes the top-of-funnel snapshot in each sponsor's post-event PDF.

Questions we get

How often does the dashboard refresh?
Every 30 seconds by default. There's a manual refresh button and a pause toggle so you can freeze the view while briefing a volunteer.
Do I need WebSockets or a background worker?
No. This uses plain 30-second polling of a compact JSON endpoint. Adding Reverb / Pusher for sub-second updates is straightforward when you actually need it — the current cadence handles queue and session decisions fine.
What powers the AI narrative?
Your configured AI provider chain — Claude, OpenAI, Gemini or Groq. The narrative is regenerated every 60 seconds (cached in between polls) to keep cost low. If every provider fails, a hand-written heuristic fallback still gives a useful summary.
What signals trigger a session capacity alert?
"High" means 80%+ confirmed RSVPs against capacity. "Full" means at or above capacity. Both show on the dashboard so you can flip the room, add chairs, or turn people away with a heads-up rather than a scene.
Can I share this view with a co-organiser?
Yes. Any tenant admin or staff role sees the same dashboard at the same URL. Bookmark it on a big TV in the ops room.
Stop running events on WhatsApp.
See the live-ops dashboard + AI operator in 5 minutes.