Opening a decision: the detail slider

What this is

Click anywhere on a decision tile's body, not one of its buttons, and the board dims behind a glass card that slides in with the row's full detail: a copper accent strip across the top, a header naming where you are, and a scrollable pane underneath. This is the detail slider. It's the same overlay component Verinode uses for a decision on the operator side of the platform, reused here on HQ Decisions so a network-level row opens in the same glass, same motion, same navigation an operator gets on their own board. What's different underneath is invisible from the outside: an HQ row's id doesn't point at a signal the way an operator's does, so a separate resolver, getHqDecisionDetail, has to work out which of Verinode's four network-level tables the row actually lives in before it can hand the pane anything to show. This article covers both halves: the pane itself, and how it finds what to put in it.

Per Verinode's privacy boundary, this is a franchisor-facing pane over franchisor-facing (and network-level) records. It reads the four tables HQ's own admins and Verinode's aggregator maintain, decision plans, interventions, consent requests, and program audits and violations, never a franchisee's own job files, invoices, or claims.

Where to find it

There's no separate URL for this pane, it's part of HQ Decisions at hq.verinode.ai/decisions. Filter the board however you're working (by status, by business area, by a search term), then click the body of any tile. The pane opens over the gallery; the filters you had set stay exactly where they were underneath it.

Opening a decision

A click on a tile's body opens the pane read-only, there's no status change from opening one. If you meant to click a button instead, Act, Park, Ignore, or Discuss, those live on the tile itself and work the same whether the pane is open or closed; see HQ Decisions: the franchisor inbox for what each one does per source. This pane is for reading a row closely before you decide what to do with it, not for changing it.

Anatomy of the pane

Header. On the left, a small uppercase "Decisions" label, then, once the row has loaded, a forward slash and the row's title. On the right, a position counter reading "N of M" and a close button (an X). Before the row finishes loading, the header shows just "Decisions," no title yet.

Body. Scrolls independently of the header, so a long row (a decision plan with a lot of background, for instance) never pushes the header or the navigation controls out of view. Inside it sits the same workspace surface an operator's own decision opens into: a status pill, the row's title, its entity name where it has one, the date it was first detected, flagged, requested, or conducted, and a short Verinode Synthesis paragraph that weaves those pieces into one plain-language read. What's deliberately absent, and why, is covered below.

Floating arrows. A left chevron and a right chevron float just outside the card's edges. The left one only appears when there's a previous row to go to; the right one only appears when there's a next one. Neither shows at the start or end of your list.

Moving through the list

The pane doesn't just show one row, it's a rail through however many rows matched the filter you had set when you opened it. Three ways to move:

• Click the floating arrows. Left for the previous row, right for the next.
• Arrow keys. Left and right arrow do the same thing, ignored while your cursor is in a 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, even while you're mid-swipe. Clicking the dimmed area outside the card does the same, as does the X in the header.

The position counter always reflects the filtered list you had open when you clicked in, not the full board. Narrow to Pending and Risk first, for example, and the counter (and the swipe order) only ever covers your open interventions. If you act on a row from its tile while the pane is open on something else, or a row you're viewing stops matching your filter for some other reason, the pane doesn't lose it: it keeps that row in the rail so you can keep reading it, but a row that's fallen out of the filter has nothing "before" it anymore, so swiping back from there is disabled. Swiping forward carries you into the rest of your filtered list as normal.

Every move to a new row triggers a fresh fetch, the pane doesn't carry the previous row's content forward while the next one loads.

Loading and error states

While a row'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 row is fetched only when you land on it.

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

Could not load this decision.

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

Behind the pane: how getHqDecisionDetail resolves a row

On the operator side, opening a decision reads one place: your operator data, the operator's own table of detected issues. HQ's Decisions board has no equivalent single table. Per The four HQ decision sources and how they map into one log, every row on the board is really one of four separate network-level records: a decision plan, an intervention, a consent request, or a program audit or violation (audits and violations are two different tables, but the mapper treats them as one combined source).

A tile's id is a UUID generated independently inside whichever of those tables the row actually came from, so the pane can't know in advance where to look. getHqDecisionDetail (the function behind this pane) solves that by checking all of them at once, in parallel, rather than guessing: it queries the network data, and the network data simultaneously, and whichever one actually returns a row wins. The other queries come back empty and are discarded. Because only one of them ever matches a given id, running all of them in parallel is cheap, and it means the pane never needs a hint about which table a row lives in threaded through the click.

Whatever row is found gets translated into the same shape the operator-side pane uses. That's why this pane looks structurally identical to the one an operator opens on their own /decisions page (the same status pill styling, the same header layout, the same Verinode Synthesis paragraph), even though the data behind it comes from an entirely different set of tables.

What's there, and what isn't

A your operator data row can carry an AI-drafted recommendation, a calculated peer-benchmark comparison, a step-by-step agent-built plan, a modeled dollar-impact figure, and an outcome record once it's closed out. None of Verinode's four network-level tables carry any of that, a decision plan, an intervention, a consent request, and a program audit or violation are franchisor records, not AI-analyzed signals, so the resolver fills those fields with nothing rather than guessing at a number. Concretely, inside this pane that means:

• No dollar figure in the workspace header, and no "Calculated" or "Estimated" tag underneath it. A network-level record doesn't carry a modeled dollar impact.
• No Your Position or Alternatives section. Those read a peer-benchmark comparison, your value against a peer average, a percentile, a distribution, that only ever gets attached to an operator's own signal. See Your Position and Alternatives for what that section looks like when it does have data to show.
• No structured, step-by-step plan panel. That panel walks through the plan IQ's agent drafts once an operator acts on a signal. None of HQ's four sources run through that agent flow. See The structured plan for the operator-side version.

What does render, on every row regardless of which of the four sources it came from: the status pill, in the same Pending, Acted, Parked, Ignored, or Resolved language used everywhere else on the platform, the row's title, its entity name where it has one (a program name, or an anonymized franchisee, see below), the date it was first detected, flagged, requested, or conducted, and the Verinode Synthesis paragraph, a short AI-written read that weaves the title, entity, and status into one paragraph, generated fresh regardless of source. For the full tour of everything this shared pane can show when it does have benchmark and plan data behind it, an operator's own view of it, see The decision card.

The privacy boundary carries through

Two of the four sources, interventions and consent requests, name a specific franchisee. That name is anonymized inside this pane the same way it is on the tile you clicked to get here: unless your network's entity model is configured as a single legal entity running every location, the pane shows a stable label built from the location's own record, for example "Franchisee #A1B2," rather than the real name. Decision plans, program audits, and program violations never carry a franchisee identity at all in the first place, opening one of those shows a program name or no entity at all, never a franchisee's own business detail. Full detail on the anonymization rule and when a network sees real names instead: The four HQ decision sources and how they map into one log.

How to use it

1. 1From hq.verinode.ai/decisions, filter the board however you're working: by status, by business area, or with a search term.
2. 2Click a tile's body (not one of its buttons) to open it in the pane.
3. 3Read the status, the title, the entity if there is one, and the Verinode Synthesis paragraph at the top.
4. 4Move to the next or previous row with the floating arrows, the arrow keys, or a swipe, without closing the pane. The counter in the header tracks your place in the filtered list you opened from.
5. 5Close with the X, Escape, or by clicking outside the card, then use the tile's own Act, Park, or Ignore button in the gallery if the row needs to move forward.

Related reading

• HQ Decisions: the franchisor inbox
• The four HQ decision sources and how they map into one log
• Decision plans in the log
• Consent requests in Decisions
• The interventions queue for at-risk locations
• Audits and violations: enforcing a program
• The decision card
• Your Position and Alternatives
• The structured plan
• What HQ sees

Data sources