Inbox Triage Agent
Sorts the inbox and writes the reply. You press send
5 of 8 emails need attention.
8 emails checked. 5 flagged. The largest gap is 47 days.
Start with Invoice query. Metro Rail has the most flagged.
What a message needs as a reply, and whether it matters, is for you; the list only watches the clock.
Why rows were flagged
- Unanswered and past the reply target40%
- Unanswered, under 4 hours left40%
- Answered after the reply target20%
Value flagged by Sender
- Metro Rail1
- Harbour Co1
- Acme1
- City Council1
- Delta1
Items to act on, worst first
| Item | Sender | Value | Hours late or left | Why |
|---|---|---|---|---|
| Invoice query | Acme | 2 | 22 | Unanswered and past the reply target |
| Newsletter reply | Delta | 1 | 24 | Unanswered and past the reply target |
| Support ticket | Metro Rail | 3 | 47 | Answered after the reply target |
| Contract change | Harbour Co | 3 | 2 | Unanswered, under 4 hours left |
| Renewal notice | City Council | 2 | 1 | Unanswered, under 4 hours left |
What to do this week
- Resolve Invoice query first: it is the most urgent on the list.
- Fix the data or the status on every flagged row, then re-run.
- Re-run this check every time the list changes, before it reaches a decision.
Do
- Fix the highest-risk item first, then run the check again on fresh data.
Don't
- Don't act on one row alone: open the source record first.
How to improve
- Add more detail per row (owner, date, source) so the next report can trace each item.
- Add a second period of data so the next report can show whether things are improving.
Demo
Overview
Most of an inbox is four or five kinds of message asking the same things. Sorting them and writing the obvious reply is an hour a day that produces nothing.
This agent reads each message, says what it is and what it needs, flags the ones that are genuinely urgent, and writes a draft reply grounded in your published documents rather than in guesses about your business. Where it has no basis for an answer, it drafts nothing and says why.
It cannot send. Auto-responders get switched off because one confident wrong reply costs more than a week of saved typing; a draft costs four seconds to approve and removes that failure entirely. It also never decides anything on your behalf — no refunds granted, no deadlines agreed, no prices quoted that are not written down.
Benefits
An hour a day back
The repetitive replies arrive written; you read and send.
Nothing important buried
Urgency is decided by content, not by how recently it arrived.
Safe to leave running
The worst case is a draft you delete.
What it does
Sorts by what it needs
Category, urgency, and the action required — not a sentiment score nobody acts on.
Drafts from your documents
Replies quote your published policies and pricing rather than inventing plausible ones.
Writes nothing it cannot ground
No basis for an answer means no draft, and a stated reason. That gap is the useful signal.
Finds the buried urgent one
The cancellation threat sitting under forty newsletters, surfaced before it becomes a churn.
Cannot send
By design and without exception. A person approves every message that leaves.
Matches how you write
Give it past replies and drafts follow your tone, not a generic corporate register.
What this changes
What this changes about your week
- It takes over
- Sorting the inbox and writing the replies that write themselves.
- Done today by
- Everyone, first thing, before real work starts.
- How often
- Daily.
- Typical manual effort
- An hour a day that produces nothing on its own.
Why Operations in particular. It drafts and cannot send. Auto-responders get switched off because one wrong reply costs more than a week of saved typing; a draft costs four seconds to approve.
Runs unattended
Started by you or by an event, and it finishes on its own. Nothing waits for someone to be at a desk.
The same standard every time
The two-hundredth item is held to the bar the first one was. Consistency is the part people cannot sustain.
It cannot act on its own
Inbox Triage Agent has no path to sending, spending or committing. That limit is why its output is safe to act on.
How it runs
This agent runs server-side through the PROMIVO runtime. Each run is logged step by step and every tool call is permission-checked before it executes.
- 1InputValidated against the agent input schema
- 2Modelclaude-sonnet-5 with automatic fallback
- 3Toolssearch_knowledge, extract_structured_data
- 4PermissionsRuns READ actions automatically; anything higher waits for approval
- 5OutputStructured result matching the output schema
- 6LoggingTokens, duration and every step recorded
Requirements
Integrations
No integrations required.
Before you start
- The messages to triage, as text
- Optionally, a knowledge base holding the policies your replies should quote
- Requires the Starter plan or higher
Use cases
Shared support inbox
Sort the queue and pre-write the routine half of it.
Sales enquiries
Every enquiry answered the same day, in your own words.
A backlog after time away
Four hundred messages sorted by what actually needs you.
Pricing
Starter
Billed monthly through your PROMIVO subscription. Cancel at any time.
Runs consume your plan allowance for agent executions and tokens. See plan limits
FAQ
Will it send email?
No, and it has no ability to. It produces drafts and a triage list. Sending stays with a person, which is what makes it safe to leave on.
How does it know our policies?
It reads the documents you published — refund terms, pricing, delivery times. Where they do not cover a question, it drafts nothing rather than improvising.
Can it handle other languages?
Yes. It drafts in the language the message was written in.
Reviews
Reviews open once customers have run this agent.
Need this agent to work differently?
Tell us what to change and our team will scope a customised version for your business.
Customize this agent