Opening an action: the slider workspace and Discuss

What this is

Click a bar on the Gantt timeline, a pill on the Calendar, or Open Full Plan from a day's task card, and Action Plans doesn't take you to a new page. Instead the board dims behind a glass card that slides in over it: a copper accent strip across the top, a small header, and a scrollable pane underneath carrying that one initiative's full detail. This is the same detail workspace Verinode uses everywhere on the platform, the identical component behind a decision tile on HQ Decisions and behind an operator's own decision on IQ. Opening an item here doesn't change anything about it: it's a closer read of a plan, an intervention, a consent request, or a program action, not a status change.

The same click that opens the pane also hands that item to your AI agent panel on the right, so you can ask about the specific thing you're looking at instead of the network in general. That handoff, and what does and doesn't show up inside the pane, is what this article covers.

Per Verinode's privacy boundary, this pane reads only the four network-level tables HQ's own admins and Verinode's aggregator maintain: decision plans, interventions, consent requests, and program audits and violations. It never opens into a franchisee's own job files, invoices, or claims, whichever of Action Plans' two views you clicked from.

Where to find it

Open Action Plans from the HQ sidebar at hq.verinode.ai/actions. Two views feed the same pane:

• Gantt (the default view). Click anywhere on a row's label cell, or on its timeline bar, and the pane opens for that initiative. See The Gantt timeline for how the board itself draws each row.
• Calendar. Click a day's pill to open that day's task list, then click a row in that list to open its task detail card, then Open Full Plan to open this pane. See The Calendar view for the day-overlay and task-card steps in between.

Both paths land you in the identical pane described below. There's no separate URL for it: it opens over whichever view and filter you had active, and the board underneath stays exactly as you left it.

Anatomy of the pane

Header. On the left, a small uppercase label that always reads "Decisions," even though you opened this from Action Plans: the header is shared code, and that eyebrow is a fixed string rather than something that changes with the page you came from. Once the item has loaded, a forward slash follows it, then the item's title. On the right, a position counter reading "N of M" and a close button (an X). Before the item finishes loading, the header shows just "Decisions," no title yet.

Body. Scrolls independently of the header, so a long initiative never pushes the header or the navigation arrows out of view. Inside it: a status pill, the item's title, its entity name where it has one (an anonymized franchisee label, or a program name), the date it was first detected, flagged, requested, or conducted, and a short Verinode Synthesis paragraph that reads those pieces together in plain language.

Because Action Plans only ever shows items that are already actioned or resolved (that's the filter that defines the whole page), the status pill you see here reads one of exactly two things: Acted (green dot) for anything still in flight, or Resolved (teal dot) for anything closed out. You will never open an item from Action Plans and see Pending, Parked, or Ignored on this pane, those statuses belong to items still sitting on Decisions.

Footer. At the bottom of the scrollable body, a plain Back to Decisions text link. It's a second way out besides the X, and it takes you to the full inbox rather than back to Action Plans specifically.

Floating arrows. A left chevron and a right chevron sit just outside the card's edges. The left one only appears when there's a previous item to move to; the right one only appears when there's a next one.

Moving through the filtered list

The pane isn't a single fixed page, it's a rail through however many items matched whatever you had filtered Action Plans to when you clicked in. Three ways to move without closing the pane:

• Click the floating arrows. Left for the previous item, right for the next.
• Arrow keys. Left and right arrow do the same thing, ignored while your cursor is in the search box or another text field so filtering the board underneath still works normally.
• Swipe. On a touchscreen, a clear horizontal drag moves you forward or back; a mostly-vertical drag (scrolling the pane's own content) doesn't trigger navigation.

Escape closes the pane outright, from anywhere. Clicking the dimmed area outside the card does the same, as does the X in the header.

The position counter, and the order you move through with the arrows or a swipe, always reflects whatever urgency bucket and search term you had set on Action Plans when you opened the item, not the whole board. Narrow to Overdue first, for example, and the counter and the swipe order only ever cover your overdue items. Every move to a new item triggers a fresh fetch: the pane doesn't carry the previous item's content forward while the next one loads.

Loading and error states

While an item's detail is being fetched, the body shows a plain centered spinner in place of any content. This happens on the very first open and again on every arrow, swipe, or key press, since each item is fetched only when you land on it.

Unlike the pane an operator gets on their own IQ decisions, this one does not poll for fresh data while it sits open. HQ's four source tables don't run a background agent that could change an item's state mid-session the way an operator's plan generation does, so the pane fetches once per item and leaves it there until you move to a different one or reopen it.

If the fetch comes back with nothing, most often because the item's id no longer matches anything in any of the four source tables (it was deleted after its row was already showing on the board), the body reads:

Could not load this decision.

Closing the pane and refreshing the board clears this; a deleted item's row won't reappear.

What doesn't carry over from the board

This is the one thing that's easy to expect and doesn't happen: the rich, color-segmented step timeline you see on the Gantt bar, or the due-date pills on the Calendar, don't reappear inside this pane as a checklist. Those visuals are a display construction Verinode builds fresh for the board itself, a synthesized read of each item's real timestamps (when a plan was drafted and activated, when an intervention was flagged and resolved, and so on), purely so the timeline and calendar have something real to draw. Opening the same item in this pane shows you the status pill, the title, the entity, the date, and the synthesis paragraph, nothing more, no step-by-step breakdown of what you just saw plotted on the bar.

That's because none of HQ's four network tables run through the same agent-drafted plan flow an operator's own decision does. A your operator data row with ordered, assignable subtasks only ever gets created for an operator acting on their own signal. A decision plan, an intervention, a consent request, and a program audit or violation are franchisor records, not AI-analyzed signals, so this pane has no structured, step-by-step plan panel to show, on Action Plans any more than it does on Decisions. See The structured action plan for what that panel looks like on the operator side, where it does exist.

For the same reason, there's no Act, Not now, or Ignore row inside this pane. Partly that's architectural (HQ's four sources move through their own admin-gated functions, not the signal-based Act/Park/Ignore an operator's workspace calls), and partly it's simply moot here: every item you can reach from Action Plans is already Acted or Resolved by definition of the page, so there's nothing left to decide from a pending state. To advance, pause, or close out a decision plan, intervention, consent request, audit, or violation, close this pane and use its entry on Decisions, where the status-changing controls live. See Act, Park, and Ignore per source kind for exactly what each button does for each of the four sources.

Handing the item to Discuss

There's no separate Discuss button to look for inside this pane, or anywhere on Action Plans. Opening an item is Discuss: the same click that slides the pane in also hands that item to your AI agent panel on the right, so its description and suggested prompts switch to match what you just opened, and if the panel was collapsed it opens automatically.

Behind the scenes, the item's signal type (something like hq_decision_plan, hq_intervention, hq_consent_request, hq_program_audit, or hq_program_violation) is translated into one of five plain tokens the moment you click in, and that token decides which prompts the agent offers and which network tables it reads to ground the conversation. A decision plan gets prompts about adoption; an intervention gets prompts about the flagged franchisee; a consent request, a program audit, and a program violation each get their own set. The full breakdown, exactly which prompts appear for each kind and exactly what data grounds the answer, is covered in "Discuss": taking a decision to the HQ agent. Everything there applies the same way whether you opened the item from a Decisions tile or from an Action Plans row: the agent panel doesn't know or care which page you clicked from, only which item you're looking at.

Verinode never takes the action for you. Discuss grounds the conversation about the item in front of you; whether to advance, pause, or close it out is still your call, made from its entry on Decisions.

The privacy boundary carries through

Interventions and consent requests are the two of the four sources that name a specific franchisee. That name is anonymized in this pane exactly as it is everywhere else on Action Plans and Decisions: unless your network's entity model is configured as a single legal entity running every location, you see a stable label built from the location's own record, for example "Franchisee #A1B2," never the franchisee's real business name. Decision plans and program audits or violations are about the network or a program, not about any one member, so they never carry a franchisee identity at all. Whatever the agent panel says back to you about the item follows the same rule: real names only surface for networks configured as a single legal entity, everyone else sees the anonymized label in the prompts and in the grounding text the agent reads.

This pane, on Action Plans exactly as on Decisions, never reaches into a franchisee's own job files, invoices, or claims. It's built entirely from the network-level tables HQ's own admins and Verinode's aggregator maintain.

How to use it

1. 1From hq.verinode.ai/actions, narrow the urgency dropdown to the bucket you're working, Overdue and Due today first, then filter or search as needed.
2. 2Click a row's label or bar in Gantt, or a pill and then Open Full Plan in Calendar, to open the pane.
3. 3Read the status pill, the title, the entity if there is one, and the Verinode Synthesis paragraph.
4. 4Ask your agent panel about it directly, it's already bound to this item the moment the pane opened, no separate Discuss click needed.
5. 5Move to the next item with the floating arrows, the arrow keys, or a swipe, without closing the pane.
6. 6Close with the X, Escape, or by clicking outside the card. If the item needs to move forward, close, or pause, make that call from its entry on Decisions instead.

Related reading

• Action Plans: the network work in flight
• Action Plans vs Decisions: how the two HQ surfaces differ
• The Gantt timeline: reading plans on the 30-day ruler
• The Calendar view: what's due each day
• Opening a decision: the detail slider
• "Discuss": taking a decision to the HQ agent
• Act, Park, and Ignore per source kind
• HQ Decisions: the franchisor inbox
• The structured action plan
• What HQ sees

Data sources