Intelligent Request Status

An agent that sweeps your environment every hour — even when you’re asleep — and tells you which requests need your attention before they become escalations.

The problem

Procurement teams spend their days reacting. A stakeholder messages asking where their PO is. A supplier emails about a stalled onboarding. A leader asks “what’s the status of the renewal we discussed?”. By the time you find out something needs you, it’s already an escalation — and you’re firefighting from the back foot.

The information you need to stay ahead is technically all in Omnea. It’s just sometimes buried — one request at a time, behind a click, in a sub-tab, in a comment thread from last week. It’s difficult to make time to manually sweep every active request every day to spot the ones drifting off course.

Omnea’s Intelligent Request Status

The AI Request Status turns Omnea from a system of record into a system of attention — directing you to the requests that need you, before they ask you.

The Agent is an always-on AI that does that sweep for you. It continuously reads the state of every active request in your environment — overdue steps, unassigned approvers, errored integrations, rejected tasks, recent comment activity — and rolls each one up into a single colored pill in your Requests table that tells you, at a glance, what needs you.

Sort by the column to see the most urgent rows first. Filter by pill state to slice your queue (“show me only the Needs attention requests”).

It will also tell you whats happening and a recommendation of what to do.

Hover any AI-set pill (the ones with the ✨ sparkle) for a short summary of what’s happening and a recommended next action.

How it works

What the Agent watches

When something meaningful happens on a request, the Agent queues a refresh:

• A new comment is posted

• A form submission is created

• A task is completed, rejected, or its deadline lapses

• A request step completes

• A request is paused or resumed

A burst of activity is collated into a single refresh — usually within the hour, though blocking events like a rejection refresh sooner.

What the Agent reads

For each refresh, the Agent loads:

• Every active step on the request, including its SLA, how overdue it is, and who’s assigned to it

• All open tasks within those steps — what kind, which form they’re attached to, whether they have unresolved errors, whether they’ve been rejected

• The latest comments on the request and who posted them

What the Agent decides

The Agent classifies the request into one of the four AI-driven pill states (On track, Slow, Needs attention, Blocked) using a deterministic precedence:

1. Blocked wins if the request is paused, or there’s an unresolved task error or a rejected task

2. Needs attention wins if an active step is unassigned, there’s no sign the request is moving on its own, or a recent comment raises a concern

3. Slow if a step is past SLA but recent activity shows it’s already moving again

4. Otherwise On track

It also writes a short factual summary of what the request is currently waiting on, and a one-sentence recommendation for the most useful next action — both shown when you hover the pill.

Setup

1. Reach out to your AM to confirm rollout availability for your environment

2. Once enabled, head to Settings → AI and toggle on Operational Status

3. The Request Management Agent immediately starts sweeping. The first generation per request happens after the next qualifying event (a comment, a task completion, etc.) — you’ll see most active requests light up within a couple of hours

What’s coming next

• Daily summary notifications for the key Requests you own

• Bulk actions from the Requests table for Needs attention rows