Common Thread
Fourteen things break at once? Find the one shared thread, and clear the innocent.
The job
Spots when complaints across many services or partners are really one failure in something they all share. It raises one incident at the source instead of fourteen, tells the others they are not at fault, puts a cost on the failure and checks recovery.
The moment
Fourteen services on the monitoring screen turn amber together, then merge into one card naming the probable cause: the national SMS gateway, with 91% confidence based on 2,310 complaints.
A meter shows 4,200 citizen hours lost so far.
One tap opens a single incident with the platform and tells all 14 service owners "stand down, it's not you".
By the next morning every tile is back to grey and marked as over. The month's ledger puts the gateway at the top of the claim for SLA credits.
What it does
- Open one platform incident
- Stand down service owners / tell partners 'it's not you'
- Send the vendor one evidence pack
- Draft and deliver the holding line for the AI agents
- Correct partner attribution for the window
What you see
A radar sweep with services on the outer ring and shared components at the centre, arcs that light as symptoms converge, and a thread card that pulls the strands into one node; grey by default
What it moves
The hours from complaints first spiking together to one incident at the source, the number of services or partners wrongly blamed each time, and citizen or customer hours lost per shared component each month.
Built for
- Product & digital
- QA, training & knowledge
- Partners & ecosystem
