Jev for support teams: what it can do and what it costs

Jev, TypeSafe AI's decision model, answers a question about text with a value your software can use: one option from a list, a score, or a yes/no probability. That suits the small, repeated decisions in the middle of this work.

These are the jobs on this site that fit support teams best. Each has its own page with the questions to ask, the pitfalls and a worked example.

• Send each ticket to the right queue and flag frustrated customers. Routing a ticket to a queue is a textbook Choice question, and TypeSafe's own docs use this example.
• Sort mail into categories, rate urgency and flag what needs a reply. Email triage is the most widely reproduced Jev workload: one request per email can assign a category, rate urgency and flag whether a human needs to see it.
• Check every message and reply for jailbreaks and hazards before the LLM acts on it. One Jev request per message can check it against several hazards at once, such as jailbreaks, requests for harm, medical advice and self-harm, and rate how much harm answering would do.

• Jev for support ticket routing
• Jev for email triage
• Jev for LLM guardrails

Jev lists $0.042 per million input tokens with output free, so the cost depends on how much text you send and how many questions you ask about it. At the typical volumes the use-case pages assume, the calculator estimates:

• Tickets (support ticket routing): about 2.1¢ per 1,000, at 3 questions and about 300 tokens each.
• Emails (email triage): about 2.4¢ per 1,000, at 3 questions and about 400 tokens each.
• Messages (LLM guardrails): about 2.3¢ per 1,000, at 5 questions and about 300 tokens each.

• Cost calculator
• Jev pricing and real runs

A plain way to begin, in order:

• Pick the one decision that costs you the most time today, usually which team gets the ticket.
• Write the options as your real queues and keep them mutually exclusive.
• Run a few hundred closed tickets through it and compare Jev's pick with the queue each ticket was finally closed in.
• Only then let it assign tickets, and keep a triage queue for the low-confidence ones.

• How to use Jev

A decision model only chooses; it does not write the reply. If you want the customer answered, pair it with a person or a language model and let Jev decide who or what gets the ticket.

From the use-case pages, Jev is not the right choice when:

• You have a single support queue, so there's nothing to route.
• You want replies drafted. Jev can't generate text; pair it with a language model for the emails that need one.
• You need the reason a message was blocked written out for the user. Jev returns probabilities, not explanations.

• What Jev can't do

Support is the cleanest first use for most teams, because you already have the labelled data: every closed ticket tells you which queue was right. That makes it easy to test Jev on your own tickets before you trust it, which is the step most teams skip.

The figures above are list-price estimates, not measured bills, and the accuracy claims on the linked pages are the published runs' own. Test Jev on a labelled sample of your own data before you rely on it.

• What is a decision model?