Take Action: safety decisions and activation
Every section in Verinode opens the same way: a hero band, then a row called **Take Action**, then an **Explore** row of metric tiles, then **Most recent**. On the Safety page (`iq.verinode.ai/safe…
On this page
What the Take Action row is
Every section in Verinode opens the same way: a hero band, then a row called Take Action, then an Explore row of metric tiles, then Most recent. On the Safety page (iq.verinode.ai/safety, "Safety" in the sidebar), Take Action is the second row you meet, right under the hero band's "Days Without Recordable" headline. It is the working surface for safety: one tile to start a conversation with IQ, one tile that tells you what to upload to switch the section on, four tiles that launch real safety tools, and then whatever decisions Verinode has actually found in your incidents, certifications, policies, and insurance.
Nothing in this row is decorative. Each tile either opens a conversation, opens a tool, or opens a decision that Verinode is recommending you act on. Verinode never decides anything on your behalf here, it surfaces what it found and you choose what to do with it.
The tiles, left to right
1. The agent activation tile ("Start here")
The first tile in the row is a dark, copper-accented cover tile labeled Start here, headlined Talk to IQ. Its subtext reads:
"I watch incidents, JHAs, certs, and policies, and warn before something hurts a job or a person."
Clicking it opens IQ in the right-hand agent panel with a seed message already written and the input box ready for your reply. The seed message IQ opens with:
"I'm IQ. I watch incidents, JHAs, certs, COIs, and policies, and surface the patterns that hurt jobs (or people) before the next one happens. Right now I have nothing to work with. The fastest way in: forward an incident report or drop a recent JHA in this thread. Once anything lands I'll start mapping recurrence patterns, flag missing certs against the jobs you're running, and draft the next JHA from your own template. What's the most recent incident or near-miss you'd want me to read first?"
This tile is a one-time on-ramp, not a permanent fixture. Once you have engaged with the Safety agent panel (any reply sent, any tile clicked that opens the panel for this section), it retires itself and does not come back. Verinode checks this both from a local flag on your device and from the server (has any conversation actually touched Safety), so it also disappears if you engaged with IQ on Safety from a different device or a prior session.
2. The unlock-section tile ("Make Safety Work" / "Deepen Safety")
The second tile tracks how much of the Safety section is actually switched on, based on which document types have flowed in. It has two states:
- Cold (nothing has flowed in yet): eyebrow Make Safety Work, headline Upload Your Data To Switch It On, and a promise line: "Recordable tracking and policy currency, the moment records land." The button reads Add Data.
- Partial (some inputs present, others missing): eyebrow Deepen Safety, headline N Of 2 Sources In. Add The Rest., and a checklist of the two inputs Safety looks for:
- Incident Reports, sourced from Incident Report PDFs, unlocking "Recordable rate and trend signals." - Safety Policies, sourced from Safety Policy Documents, unlocking "Policy currency and coverage checks."
Each input that has already landed shows a filled green check and reads "· in"; each missing input shows an open copper ring with its label and source tool. The button reads Add What's Missing once anything is present. Either button fires the same capture flow as the Add Data button in the page header, so there is no separate upload path to learn.
Once both incident reports and safety policy documents have flowed in, this tile stops rendering entirely.
Note
Safety's cold state is one of the sections with its own rich empty-state card already built into the Decisions row below (see "Build your safety record" further down), so on web the unlock tile stays hidden while Safety is fully cold and only appears once you're partway in, to nudge you toward the input you're still missing. On mobile it owns the cold state directly, since the mobile empty states don't carry the same upload guidance.
3. The four launch decks
Next come four launch tiles, each a dark album-cover tile with a breathing "Live" dot in the corner and a plain-language call to action. Clicking any of them opens that tool as a center overlay, full navigation and the IQ panel stay in place around it.
Safety ("Track incidents and your safety record", Start). This is the incident logging tool. Its Overview tab explains the value in one line: log every incident and near-miss for a clean OSHA record and a safer crew, in under a minute. Its Open items tab lists incidents that are not yet closed, showing the incident type and its status and date. Its Report an incident tab is the flat intake form. The primary action from this deck is Report an incident.
Corrective Actions ("Close the loop on every incident", Start). This is a standalone workflow over the fix-and-due-date fields that already live on every incident, separate from logging the incident itself. Its Overview explains: set a fix and a due date on every incident so nothing slips. Its Open actions tab lists incidents that have a corrective action set but not yet closed, each showing the incident type and its due date (or "not set"). Its Set an action tab is where you attach a corrective action and due date to an open incident. The primary action is Set a corrective action.
Safety Policies ("Write the rules your crew follows", Generate). This is a document generator. Its Overview frames the value directly: put your safety rules in writing, because a written, acknowledged policy is what an OSHA inspector and a carrier both want to see. Its Library tab lists every policy you have written, with its category and version number (e.g. "v2"). Its Write a policy tab is the generator panel. The primary action is Write a policy.
Emergency Plans ("Plan for fires, weather, and spills", Generate). Another document generator, this one for site-level emergency response: muster points, chain of command, who calls whom. Its Overview: have a plan before you need it, because fires, severe weather, and chemical spills go better with a written plan the crew already knows. Its Library tab lists plans by type and site. Its Write a plan tab is the generator. The primary action is Write a plan.
All four decks show a realistic example (a sample safety scorecard or a sample policy document) with a caption clarifying it is a worked example, not your data, until you have logged enough for your own numbers to fill in.
4. The SafetyDecisionsRow
After the four launch tiles, the row keeps going with whatever real decisions Verinode has found, or one of three honest empty-state cards when it has not. This row loads on its own, in the background, right when the Safety page opens.
When Verinode has found decisions. Each one renders as a decision tile: a copper Recommended label, a headline that is either a dollar figure (with "/mo" or the relevant period beside it) when the decision has a quantified cost, or the metric's plain-language name when it does not. A subtext line names what the number is about (for example, which safety pattern is driving it) whenever the headline is a dollar figure. Below that sits the entity the decision concerns (an incident, a policy, a vendor, whatever the signal traces back to), a four-stage lifecycle indicator (Flagged → Planned → Acting → Resolved) showing how far along you are in acting on it, and, when Verinode is tracking a trend, an arrow and a label reading Improving, Declining, or Stable. Clicking a tile opens the full decision detail as an overlay on top of the page. The first tile in the row is shown larger (a "hero" variant); the rest are standard size. Only decisions you have not yet acted on or resolved show up here, once you have acted on or resolved one, it moves to the decision workspace instead of cluttering this row.
When there is nothing to show, one of three honest empty states appears instead of a fake or padded tile:
No records yet. Title: "Build your safety record." Body: "Three quick moves and your first safety decisions surface within minutes." This card lists three concrete steps:
- 1Upload an OSHA log or incident report. OSHA 300 / 301 forms, near-miss reports, or any internal incident write-up. Clicking opens the data-capture flow directly.
- 2Drop your insurance COI or workers' comp policy. PDFs of GL, WC, auto, umbrella, anything the carrier sent over. Clicking opens the data-capture flow directly.
- 3Forward training and cert confirmation emails. Set up auto-forward so renewal reminders flow in automatically. This step links to Connecting your data rather than opening a capture flow, since it is a one-time setup step.
A footer under the steps reads: "Operational-risk decisions surface as the detector reads your data," alongside a short trust line reminding you how Verinode handles what you send it.
Still learning. Title: "Still learning your safety record." Body: "As the detector analyzes incidents, training compliance, and peer incident-rate patterns, top decisions will appear here." You get this state once records exist (incidents, certs, policies, or insurance are on file) but the detector has not yet surfaced anything worth flagging. This is not a stall, it means your data has landed and Verinode is actively working through it; give it a pass before assuming something is wrong.
All clear. Title: "All clear on your safety." Body: "You've worked through N signals. New ones will surface here as the detector finds them, until then, nothing needs your attention." (N is the number of past signals you have resolved.) You land here after Verinode has found and you have worked through every safety signal it has raised so far, a genuinely clean state, not an empty one.
While any of this is loading, the row shows three pulsing placeholder tiles rather than a blank space, so the page never looks broken while decisions are being fetched.
How to read the row as a whole
Think of Take Action as answering, in order: how do I start talking to IQ about safety (tile 1), what data is Safety still missing (tile 2), which tools can I open right now to do something concrete (the four launch decks), and what has Verinode already found that needs my attention (the Decisions row). Most operators will use the row differently depending on where they are:
- Brand new to Safety: start with tile 1 or the "Build your safety record" steps, either forwards you straight into the capture flow.
- Some data in, section still cold or partial: use the unlock tile's checklist to see exactly which of the two inputs (incident reports, safety policies) is still missing, and add it.
- Data flowing, no decisions surfaced yet: trust the "Still learning" state. It means Verinode has records but hasn't found a pattern worth flagging yet; keep feeding it and check the Explore row's numbers (Days Without Recordable, Cert Wall, Policy Currency) in the meantime.
- Decisions present: work the hero tile first (it is ranked highest), then the rest, following the lifecycle dots to see what is flagged, planned, being acted on, or resolved.
- Everything clear: treat "All clear" as a genuinely clean bill of health, not a bug. New signals will surface here the moment the detector finds one.
Data sources
Data sources
- 1.Your incident reports, near-miss write-ups, and OSHA logs. Your business.
- 2.Your safety policy documents and their review cycles. Your business.
- 3.Your certification and training records. Your business.
- 4.Your insurance certificates (COIs) and renewal dates. Your business.
- 5.Peer incident-rate and training-compliance benchmarks. Verinode intelligence layer (anonymized, never sold to carriers).