What the AI judges
- Which team should handle it, and how urgent it is
- Whether a person, not a standard reply, must handle it
- Which prepared reply fits
- Which date is the deadline and which amount is at issue, chosen only from those found in the email
Free for UK teams: a one-page AI policy template you can adapt today.
Get the templateLive demo · AI agent
Paste a customer email and a live agent routes it, rates urgency, flags when a person must reply, picks a holding reply, and finds the deadline and amount.
Or try a sample
Your text is sent to TypeSafe to produce the result and is not stored by Nerdster. Please don't paste confidential information. Privacy
How it works
The demo runs on Jev, TypeSafe's System One model, which returns probabilities and choices rather than writing text. Every verdict comes from explicit rules applied to those judgments, which is how we build agents for clients too.
Samples are fictional and their results were precomputed with the same pipeline. Results are automated judgments to check, not advice.
NEXT STEP
This is the shape of a real agent: narrow judgments, rules in code, a person where it matters.
No. It chooses one of five prepared replies. In a real build you approve the wording, and a person handles anything that needs judgement.
No. Code finds every date, amount and reference in the email first; the model can only choose among those, or choose "none".
Please do not. Use a sample or an invented email. Text is sent to TypeSafe, the AI provider, to produce the result, and is not stored by Nerdster.
A typical run takes about a quarter of a second of model time. A production agent can also check your records before routing.
More live demos