The Action focus tab lines your actions up against your survey - by question and by category - so you can see where effort is concentrated and where it's thin. Use it to spot mismatches between where attention is going and where the survey suggests it matters most.
The filter on the right changes the data shown:
The bubble chart: "Are we focusing on the right things?"
Each bubble is one survey question:
Higher up = more important. Its position reflects how strongly that question drives your overall engagement score, based on your own survey's statistics.
Further right = more action attention. The centre line is a "fair share" — what you'd expect if effort were spread evenly.
Bigger bubble = more actions, with the count printed inside.
The four quadrants:
Quadrant | Meaning | How to read it |
Priority gap (top left) | Important but under-served | Worth a look first - check whether it's a genuine gap or already well-handled outside formal actions. |
On track (top right) | Important and well-served | Keep momentum; no need to pile on more |
Over-indexed (bottom right) | Heavy effort on lower-impact topics | Check it's a deliberate choice, not drift |
Low risk (bottom left) | Low importance, low effort | Usually fine to leave for now |
The quadrants are prompts, not verdicts - there can be good reasons a topic sits where it does. Click any bubble to open the actions behind it and see for yourself. A simple working rhythm: scan the top left first, click in, and where it's warranted, add a clear owned action - then watch the bubble move right over time.
The chart is calculated from your survey data and action links, not written by AI. Prism supplies the importance statistics; everything else is straight arithmetic, so the numbers are reproducible.
The chart only counts actions linked to a survey question. Actions created without a question attached fall into a single "Actions not linked to questions" bubble (visible in the All questions view).
If you've done lots of work on a topic but its bubble looks small, check whether those actions were linked to the question. The most reliable way to keep the chart accurate is to create actions from within the relevant question - that links them automatically.
Use the view toggle to switch between your Top 5 key drivers (default) and all questions.
Actions by question
A table with one row per survey question, showing its impact, response favourability, and its actions split by status. Use it to find a low-scoring, high-impact question with few or no actions - clear candidate for attention - use the "⚡️" button on that row to start a conversation with the relevant manager -worth checking their local context first, rather than assuming the gap needs filling.
The Category column shows each question's quadrant from the chart, so you can filter to just the priority gaps.
Dropdown selection:
Actions by category
Action counts grouped by topic - the category tag Prism assigns to each action. It shows the shape of activity by theme (e.g. more actions on communication than career development) and lets you jump to all actions on a topic.
Dropdown selection
Common questions
Why is a question in the Priority gap quadrant? It ranks in the more important half of your questions, but is getting fewer actions than an even spread would give it. In plain terms: it looks like it matters, but it isn't getting attention.
Why don't the question counts and category counts match? They count different things. "Actions by question" counts actions by the survey question they're linked to; "Actions by category" counts them by their topic tag. One action can carry several topic tags (or none), so the totals aren't expected to match.
The chart says "unavailable" / I can't see Top 5 key drivers. Key drivers come from a single survey. With "All dashboards" selected there's no single set of key drivers, so switch to one dashboard.
Why has the bubble chart disappeared entirely? It only shows when Prism is enabled for your dashboard. The two grids below it still work without Prism.
The chart is taking a while to load. It's computed in the background and the page checks for the result. If it says it couldn't finish in time, refresh the page.
Some bubbles look tightly positioned - how precise is this? Importance comes from your survey's correlation statistics, which are less stable for small response groups. Treat positioning as a guide for conversation rather than a precise score, especially for smaller teams.