Screen monitoring and automated outreach · Ireland, UK and Europe
Dark screens, found and fixed before anyone calls
Nodark watches every screen in your estate, texts the venue, phones with an AI agent, and raises a callout only when one is needed.
Hover a screen. Demonstration data.
The problem
You find out from the client
A dark screen is unsellable inventory. It earns nothing, and it is still on the plan. Most operators find out when the client rings, or when someone walks past it. By then the screen has been off for days. Your CMS knew within minutes. Nobody was watching it, because nobody has time to watch thousands of rows.
Dark screens don't bill
Every hour a screen is off is inventory you sold and cannot deliver. The loss is silent.
The client notices first
Nothing damages an account faster than a client telling you your own screen has been dark all week.
Nobody answers the phone
Venue contacts ignore unknown numbers and voicemails. The one person who can fix it is unreachable.
Callouts for nothing
Around six in ten single-screen outages clear with a router power-cycle. A €85 callout to press a switch.
Alerts nobody reads
Flat 24/7 monitoring flags every closed pub overnight. The inbox fills, the team stops looking at it.
No record afterwards
You cannot show a client what you did, when you did it, or why it took that long.
Escalation
It doesn't alert you. It handles it.
Nodark connects to your CMS read-only and checks every screen on a roughly 15-minute cycle. When one stops reporting, this runs on its own.
- 01
Detect
A screen misses its check-in. Nodark confirms the venue is open, checks it isn't a scheduled content rule, and opens an incident within minutes.
- 02
Text the venue
The site contact gets a message about that specific screen. Has the power gone? Wi-Fi changed? Building works? Replies are read and acted on.
- 03
Text again
No reply, so a second message goes out. Still inside the venue's opening hours, still inside the response window you have agreed.
- 04
Phone with an AI agent
A conversational agent calls the contact. It asks diagnostic questions, walks them through a router power-cycle, and records what it learned and why.
- 05
Flag a callout
Only if a real fault is confirmed. The callout is raised for your approval, with the diagnosis attached, before anyone is dispatched.
Your team owns escalations and dispatch. Nodark does the watching, the chasing and the diagnosis, then hands you a decision.
Watch it happen
One incident, start to finish
Detection, a hold until the agreed response window opens, one text, a reply understood, a quick fix, recovery, closed. Nobody in the office touched it.
Text thread · site contact
Incident log
- 09:41DetectedScreen goes dark. An incident opens against the client's agreed response window.
- 09:41HoldOutreach is scheduled for the response window you agreed. No premature contact.
- 09:56Text sentWindow open. Site contact texted — venue-hours and quiet-hours checks passed.
- 10:01Reply readThe reply is read and the right diagnostic branch is picked.
- 10:02Quick fixRouter power-cycle suggested. Around six in ten single-screen outages clear this way.
- 10:10RecoveredScreen reports back in. It is online.
- 10:11ResolvedIncident closed, contact thanked. No engineer needed.
A composite of a typical incident, replayed at real speed. Thirty minutes of estate time, start to finish, with nobody in the office.
One incident
73 minutes, start to finish, unattended
One incident on a 9,000-screen estate. From the screen going dark to a callout raised for approval, no person touched it.
- T+0
Screen goes dark
A screen misses its check-in. The venue is open. No content rule applies. An incident opens.
- SMS 1
First text
The site contact is asked what has changed at that screen — power, Wi-Fi, building works.
- SMS 2
Second text
No reply. A follow-up goes out, still inside opening hours and inside the agreed response window.
- Call
AI voice call
The agent phones the contact, works through the symptoms, and establishes the screen is physically fine but has lost network.
- T+73
Callout raised for approval
Hardware fault ruled out, cause identified, callout queued for a human to sign off before any cost.
Texted twice, phoned the contact, diagnosed a network fault, ruled out a hardware failure, and put a callout in front of a human in 73 minutes. Nobody on the ops team touched it.
A 21-second call that saved a visit
One AI voice call ran 21 seconds. The contact power-cycled the router while on the phone. No engineer went out.
Self-cancelling callouts
In testing, a callout opened at 17:49 and cancelled itself at 17:51 when the screen came back. Nothing was dispatched and nothing was billed.
What changes
Before and after
- Weeks in the dark
- Minutes to detection
- The client tells you
- You tell the client
- Voicemail, no answer
- A call that gets answered
- Engineers sent to guess
- Fault confirmed first
- Callouts discovered on the invoice
- Callouts approved before spend
- Alert inbox nobody reads
- A short list that needs a person
Guardrails
What it will never do
Automated contact goes to real people in real venues. The limits are hard-coded, not settings you have to remember. Outreach is switched on client by client, so nobody on your estate gets contacted before you say they are ready.
Business hours only
Calls are placed between 9am and 5pm. Never at 3am, never at the weekend by accident.
Only when the venue is open
A screen that is off in a closed pub is not a fault, and nobody gets texted about it.
Two calls a day, five a week
Per venue. A stubborn screen never turns into a nuisance caller.
No spend without a human
A callout is raised for approval. Your team approves it. Nobody is dispatched otherwise.
See it run on your estate
Give Nodark read-only access to your CMS. Within a week you will see which screens are dark, and how many clear without an engineer.