Live task routing race

The same stream of workplace requests, the same questions, two kinds of AI.

Waiting 0 arriving / s 0 watching elapsed 0:00

Jev

Jev 1.14 decision model
batches of 20
Processed0
Waiting0
Per second0
Latency-
Spend$0.00
auto-routedmanager decisionhuman review

LLM

Claude Sonnet 5.5
10 calls in parallel
Processed0
Waiting0
Per second0
Latency-
Spend$0.00
budget $5.00

Accuracy on the same items

Jev
-
LLM
-
Jev + LLM for unsure
-

Measured against the hidden true department of each request.

Cost and time

Cost per 1,000 requests--
Time to clear the queue--
Hybrid cost per 1,000-

 

Where Jev sent the work: channel, department, outcome

Live feed