Skip to content
All posts

Harness Archy Insights to spot risks and act on them

How to use the AI Assistant Insights in ArchyAssistant to surface blockers and risks in a Jira project, decide which ones matter, and turn them into backlog work your team can prioritise.

Guides5 min read

A translucent conversation panel passes through a luminous keyhole to an editorial page, illustrating a secure publishing connection.
null

Every project carries risks that nobody has written down yet: a dependency that is slipping, an area of the backlog that is thinner than it looks, a feature that matters to users but is still at an early stage. By the time these show up in a sprint review, they are already expensive. The earlier a team sees them, the cheaper they are to fix.

ArchyAssistant, the AI assistant for Jira, includes AI Assistant Insights for exactly this. It reads the state of your project and points out improvements, blockers and issues worth attention. This guide shows a simple routine for working with insights: review them, decide which ones matter, and turn the important ones into backlog work the team can prioritise.

About this guide

This guide is based on a 2025 walkthrough of ArchyAssistant (then called Archy for Jira). The current ArchyAssistant release is built on Atlassian Forge, so labels, schedules and layouts may differ from what is described here.

An insight is only useful if it ends up as a decision: act on it, plan it, or deliberately set it aside.

What insights are

Insights appear on the ArchyAssistant dashboard inside your Jira project. Each one describes something ArchyAssistant has noticed about the project, such as a gap in scope, a blocker or a part of the plan that carries more risk than the rest. In the 2025 walkthrough, insights were generated on a 48-hour cycle, so a team could treat them as a regular input to planning rather than a one-off report.

ArchyAssistant dashboard in a Jira project shown on a desktop monitor, with the Next best action and AI Assistant Insights cards

Step 1: Make reviewing insights a habit

Insights only help if someone reads them. Put a short review on the team's calendar that matches the generation cycle, for example every second day during active development. A reminder is enough. The point is consistency: a risk spotted on Tuesday and read on Friday has already cost three days.

Step 2: Review and evaluate what matters

Open the insights on the dashboard and read each one against what you know about the project. Look for improvements, blockers and issues in areas that are critical to the product. In the walkthrough's example project, a merchant finance platform, ArchyAssistant flagged that financial health scoring and the risk assessment models were still at an early stage. Both were central to the value users would get, so they deserved attention now rather than later.

Ask two questions of every insight: is this accurate for our project, and does it affect something users care about? The answers decide what happens next.

Step 3: Keep, act on or delete

Act on it

Draft actions to mitigate the issue. Note the dependencies for each action and assign an owner, so nothing waits on an unclear handover.

Ask for options

Ask ArchyAssistant to suggest further ways to address the issue when the right response is not obvious.

Set it aside

Delete insights that are not relevant to your project. Do it deliberately; an insight that looks minor can still point to something real.

Step 4: Turn insights into backlog items

For insights you decide to act on, let ArchyAssistant propose additional backlog items that address them. It suggests new epics and user stories linked to the issue it found. Review these suggestions the same way you would review any backlog proposal: check relevance, edit descriptions and remove what does not fit.

Then go one level deeper. Open each relevant epic, add user stories where coverage is thin and refine existing stories until they are clear and complete. The aim is work the team can estimate and pick up, not a list of concerns.

Step 5: Save to the backlog and prioritise

Save the resulting epics and stories to the Jira backlog, then prioritise them alongside existing work. Keep the backlog current as priorities change, so it continues to reflect what the project actually needs.

Cautions

  • Make sure every team member understands why insights are reviewed regularly. A skipped review is how critical issues get missed.

  • Respect the dependencies you recorded for each action. Ignoring them is a common cause of project delays.

  • Do not delete insights without thinking them through. They may hold information the team needs later.

Tips for a smoother routine

  • Set a recurring reminder that matches the insight cycle.

  • Discuss insights in the team's usual collaboration channel so decisions are visible.

  • Invite feedback on insights and the actions that follow, to build a habit of continuous improvement.

  • Review and update the backlog regularly so it stays aligned with project goals.

The outcome

Followed consistently, this routine turns a passive feed of observations into a steady flow of decisions. Risks surface early, the important ones become planned backlog work with owners and dependencies, and the rest are set aside on purpose rather than forgotten.

Bring ArchyAssistant into your Jira projects

See how ArchyAssistant helps teams spot risks and plan the work in Jira, or install it from the Atlassian Marketplace.

Make your next piece count

Give good ideas a better way to ship.

Start with a brief. Keep the editorial decisions yours.

Part of the Archy AI Platform