Project Health AI

By the time you notice, it's too late. So AI tells you first.

Automatic project-health scoring across five dimensions. When reply stalls, an engagement drop, or unread materials appear, the AI raises the alert.

WHY IT MATTERS

Catch the signs of a cooling client project before any human can

AI surfaces the anomalies that field instinct and experience tend to catch too late.

  1. 01

    AI catches anomalies before you do

    The five-dimension score updates daily. When thresholds are crossed, AI surfaces the warning sign and notifies the rep.

  2. 02

    See the evidence, follow up fast

    Alerts arrive with concrete evidence — "zero views from the decision-maker for 7 days," "key material unread" — so you can act immediately.

  3. 03

    Managers rescue client projects

    Managers see at-risk client projects in one list and intervene as an organization, even before the rep notices.

5-dimension score × AI alerts

Watch the numbers, let AI flag the moment it shifts

Five-dimension scoring across Activity, Engagement, Stakeholder, Momentum, and Response — paired with AI alerts that catch reply stalls, engagement drops, unread material, and fading client-side champions.

  • Daily score updates capture every trend
  • AI surfaces meaningful anomalies in real time
  • One click to jump to the underlying evidence
🔴

Project risk detected for Yamada Mfg. — copy has been awaiting a reply for 14 days

Project Health Dashboard

92
Aoba TradingAcceptance
45
Yamada Mfg.Production

Awaiting reply 14 days — resend the request

78
Wakaba LogisticsKickoff

Yamada Mfg. — risk analysis

Client engagement
32%
Stakeholder involvement
65%
Task completion
88%

Key capabilities

Automatic scoring across five dimensions

Five axes scored daily — activity, client engagement, stakeholder involvement, momentum, and response speed — each traceable back to the records behind the number.

AI alerts — reply stalls detection

Detects when access stops over a defined window. AI surfaces "no customer response" before you notice on your own.

AI alerts — sudden engagement drop

Fires an alert when recent view frequency or dwell time falls sharply. Catch the early signs of a cooling client project.

AI alerts — unread materials

Notifies you when important shared materials sit unread for too long, so nothing you send disappears into a void.

Score trend graph

Visualizes the time series of project-health scores. Track how each follow-up or proposal change shifts the score.

Customize from workspace settings

Tune the weight of each dimension and the threshold for each alert to match your project process or industry.

FAQ

Frequently asked questions

  • What data is the score calculated from?

    It combines five signals: Room access logs, file dwell time, message response speed, stakeholder involvement, and task progress.

  • Can I tune the alert thresholds?

    From workspace settings you can change thresholds for each alert (reply stalls / engagement drop / unread material / fading champion, etc.). For longer project cycles, you might set "reply stalls = 14 days," for example.

  • Won't I get too many false positives?

    Early on, detections do tend to run a little high. Settings let you filter — for example, "only Slack-notify Severity: High" and "aggregate Medium-and-below on the dashboard."

  • What can managers do with this?

    From the dashboard, managers see a unified score list across every client project. They can filter for "declining health" or "fading champion" client projects and prioritize follow-up.

  • How does this relate to AI Agent?

    Project Health handles anomaly detection; AI Agent proposes the next action given the anomaly. Together, you can automate from detection to response.

Try Terasu free for 14 days

Full Business plan features. No credit card required.

Project Health AI | Terasu