The Safety section: your operational-risk radar

What the Safety section is

Safety is where Verinode watches the operational-risk side of the business: recordable incidents, corrective actions, safety policies, and the insurance that backs the crew's work. It reads the incident reports, policy documents, and certificates of insurance you already generate, turns them into one running record, and surfaces the handful of things that actually need a decision this week, the way a fractional COO would flag them in a Monday morning walkthrough.

Verinode does not file OSHA paperwork for you and does not decide whether an incident is recordable. It reads what you send it, classifies it against the OSHA framework and your own policy cycle, and lays out the decision. You act on it.

Safety sits in the Compliance group in the sidebar, alongside Compliance and Certifications. The three share a data posture (Premier-gated, activation-gated) but each watches a different slice: Compliance watches regulatory frameworks and audits, Certifications watches individual and firm credentials, Safety watches incidents, policies, and the insurance that covers them.

Where to find it

Open Safety from the sidebar, under the Compliance group, at iq.verinode.ai/safety. Like every other section, it must be switched on before it runs (see Certifications: activation and Premier gating for how Section Activation works, the mechanism is identical across sections). Once active, the page opens straight into the Safety home, a full-height view with no separate "loading" step.

Premier-only gating

Safety is one of Verinode's Premier-tier sections, alongside Compliance, Certifications, Processes, and Advisory. What you see depends on your membership:

• Contributor membership. You see a single summary line instead of the working page. If Verinode already holds safety records for you, it reads "Verinode is tracking [N] safety records. Upgrade to Premier for the full safety intelligence layer," with the real count of records on file, incidents, policies, and insurance combined. With nothing on file yet, it reads "Safety intelligence, incident logs, OSHA radar, depth-chart coverage, unlocks at Premier."
• Executive membership. You get a working preview: your single most recent safety record shown in full, with the rest of the record set gated behind an unlock banner and an Upgrade to Premier button.
• Premier membership (and above). The full section: every incident, policy, and insurance record, the complete Explore row of operational-risk metrics, the cards-slider with all five tabs, and peer benchmarks where a cohort exists.

Whatever count Verinode shows at a lower tier is a real number pulled from your own records, never a placeholder built to make the upgrade feel more urgent than it is.

Architecture: three rows, one overlay

Safety follows the same shell every entity section on the platform uses: a fixed home behind, and a full-screen cards-slider that opens on top of it when you click into anything. The home never unmounts, closing the slider (Escape, or the back bar) drops you back exactly where you left the page.

The home itself is three rows stacked under a hero panel:

1. Hero. The headline number and two secondary stats, always visible, no click required.
2. Take Action. The agent entry point, a data-completeness nudge, four launch tiles for logging safety work directly, and your live safety decisions.
3. Explore. Six tiles, each a different cut of the operational-risk picture, each one click from a matching tab in the slider.
4. Most recent. A scrolling strip of your newest and most urgent records across every kind, sorted by severity first and due date second.

Clicking any tile in Take Action, Explore, or Most recent opens the cards-slider, a five-tab overlay (Findings, All Safety, Standing, Exposure, Benchmarks) that reads the same record set the home does. The rest of this article walks through all four rows and all five tabs.

The Hero panel

At the top of the page, one number leads:

• Days Without Recordable, the eyebrow above a large count. This is the number of days since your most recent OSHA-recordable incident. With no recordable incidents on file at all, it reads 365, a clean-slate default, not a claim that you have gone exactly a year without one.
• A pill beside the headline reads training compliance as a percentage ("X% Trained"), colored green at 90% and above, amber from 70% to 89%, and copper below 70%.
• The line under the headline explains itself: with no safety records at all it reads "Add Data, incidents, certs, policies, and COIs land here as they come in." With records on file, it names whichever is more pressing, a cert-wall count if certifications are expiring in the next 30 days, otherwise the days-without-recordable line again in prose form.
• Two secondary stats sit beside the headline: Cert Wall (the count of certifications expiring in 30 days, reads "Nothing Expiring Soon" at zero) and Insurance Due (the count of insurance policies expiring in 60 days, reads "All Policies Current" at zero).

Certification tracking itself now lives in its own Certifications section, so the Cert Wall figure and the Trained pill on this page reflect that split: today they read zero and 100% respectively on Safety's own hero, while the real credential detail, renewal calendar, and per-tech currency live on the Certifications page. Incidents, policies, and insurance are what actively drive this page's numbers.

Take Action row

The first row under the hero is where you either start a conversation or start a document:

• Ask the agent. The first tile is your entry point into a conversation with IQ about safety specifically, framed around this section's data.
• Data-completeness nudge. If Verinode has some but not all of the inputs it needs (incident report PDFs, safety policy documents), a tile names exactly what is missing and what uploading it turns on, for example, incident reports unlock recordable-rate and trend signals, policy documents unlock policy-currency and coverage checks. With everything captured, this tile disappears.
• Four launch tiles, each opening a dedicated workspace:

- Safety ("Track incidents and your safety record"), the incident log itself: an overview of your safety standing, an Open Items list of incidents still needing a corrective action, and a flat "Report an incident" form. - Corrective Actions ("Close the loop on every incident"), a workflow surface purpose-built for setting a fix and a due date on an open incident, tracking it, and marking it done, the fastest lever on your EMR. - Safety Policies ("Write the rules your crew follows"), a document generator: an example policy for reference, your policy library, and a "write a policy" flow that produces a versioned, dated document your crew acknowledges. - Emergency Plans ("Plan for fires, weather, and spills"), the same generator pattern for site evacuation and emergency-response plans.

• Your live decisions. Below the launch tiles, up to ten real decisions Verinode's detectors have raised for Safety, rendered the same way every section's decisions render (see The decision workspace).

With no decisions to show, this row is honest about why, in one of three ways:

• No safety records at all ("Build your safety record"): "Three quick moves and your first safety decisions surface within minutes," with three concrete steps, uploading an OSHA log or incident report, dropping an insurance COI or workers' comp policy, and setting up auto-forwarding for training and cert confirmation emails, plus a line reminding you that operational-risk decisions surface as the detector reads your data.
• Records exist, but the detector has not run its pass yet ("Still learning your safety record"): "As the detector analyzes incidents, training compliance, and peer incident-rate patterns, top decisions will appear here."
• You have worked through every signal that has fired ("All clear on your safety"): a note telling you how many signals you have resolved so far, with new ones surfacing as the detector finds them.

Explore row: the six operational-risk tiles

Explore is a horizontal row of six tiles, each with a headline number, a small inline chart, and, where a peer cohort exists, a delta against the peer or industry figure. Every tile opens the matching slider tab.

1. All Safety. The total record count, broken down as "N incidents · N policies · N COIs" (certifications, once they flow back through this count, will add a fourth segment). A segmented bar shows the mix. Opens All Safety.
2. Days Without Recordable. The same headline as the hero, with a gauge that fills toward 365 days and turns copper the moment an incident rate is on the books. Below it: your incident rate in incidents per 100,000 team hours, or "Clean record YTD" if nothing has been recorded this year. Opens Benchmarks, focused on the incident-rate row.

The incident rate is the standard OSHA formula: year-to-date OSHA-recordable incidents, multiplied by 200,000 (the hours 100 full-time employees work in a year), divided by your team's actual hours this year (active team headcount times 2,000 hours). A smaller number is better, more team hours behind the same incident count lowers the rate.

1. Cert Wall. How many certifications are expiring within 30 days, alongside your training compliance percentage. Opens Exposure, focused on training.
2. Policy Currency. The percentage of your active safety policies whose review date has not yet passed, with a marker at the 90% line, green at or above 90%, amber from 70% to 89%, copper below. With no active policies at all, this reads 100% by default (nothing is overdue because nothing exists to be overdue). Opens Standing, focused on policy currency.
3. Insurance Renewals. The count of insurance policies expiring within 60 days, with a two-segment bar splitting current policies from the ones coming due. Opens Exposure, focused on insurance.
4. Incident Follow-Through. The median number of days between an incident occurring and its corrective action being closed, mined directly from your own closed incidents. With nothing closed yet, it reads "Close incidents to map your follow-through time." A peer figure appears alongside it once enough of your own incidents have both dates on file.

Where a peer or industry figure exists for a tile (incident rate, training compliance, policy currency), the delta reads "vs Peer" when a live peer cohort has formed on that metric, or "vs Industry" when Verinode is comparing you to a published research baseline instead. Either way, the comparison is never sold to carriers, it exists purely so you can see where you stand.

Most recent row

The bottom row is a scrolling strip of individual records, mixing incidents, policies, and insurance in one list, most urgent first: critical severity ahead of warning, warning ahead of informational, and within each band, overdue items ahead of items with a future due date, ahead of items with no due date at all. Each tile shows a kind label, the record's title, a colored severity dot, its status, and a due badge when something is overdue ("Xd overdue") or coming due within 30 days ("Xd"). Clicking a tile opens that exact record in the All Safety tab of the slider.

With no records on file yet, the row reads: "Safety records will appear as you forward incident reports, drop cert scans, or paste insurance COIs."

The record model: one shape, several kinds

Every row Safety shows, in the Most recent strip and inside the slider, is a SafetyRecord: a single shared shape (title, subtitle, due date, status label, and a severity signal) that different underlying record kinds project themselves into. Four kinds share this shape today, though only three actively flow through the Safety page:

• Incident. Sourced from your incident log: incident date and type, severity (first aid, recordable, lost time, or fatality), location, description, root cause, corrective action and its due date, whether it is OSHA-recordable, days away from work, and status (open, investigating, corrective action, closed). The title is the incident type in plain words (for example, "slip and fall" becomes "Slip And Fall"). Severity reads warning when the incident is lost-time or OSHA-recordable, critical when it is a fatality, and critical regardless of the underlying severity if its corrective action is overdue and still open.
• Policy. Sourced from your safety-policy library: title, category, effective and review dates, version number, and status (draft, active, under review, archived). The subtitle combines category and version (for example, "fall protection · v2"). An active policy whose review date has passed shows "Review overdue Xd" and flips to warning severity.
• Insurance. Sourced from your insurance policies: policy type, carrier, policy number, coverage limit, deductible, annual premium, effective and expiration dates, whether it auto-renews, its experience modification rate (EMR), and status. The title joins policy type and carrier with a dash, for example a general liability policy from a named carrier. An expired policy is critical severity; one expiring within 30 days, or within 60 days without auto-renew, is warning. A workers' comp policy with an EMR of 1.25 or higher is always critical, regardless of its expiration date, an EMR that high is itself the risk, not the calendar.
• Certification. Still defined in the data model for compatibility, but certifications migrated to their own Certifications section. Safety's own page no longer reads or projects certification rows directly.

Drilling in: the cards-slider's five tabs

Clicking into any tile opens the slider. It always opens on the tab tied to whatever you clicked, but you can switch freely between all five:

• Findings. Open tips (short, per-record nudges Verinode's detectors have raised) at the top, followed by the same unified findings grid every section uses for its live decisions.
• All Safety. Every record in one searchable list. Filter pills across the top (All, Incidents, Certs, Policies, Insurance) each show their own count. A search box narrows by title, subtitle, or status. The list shows up to 120 rows at a time, with a note to narrow your search if there are more; with no records matching a filter or search, it reads "No records match the filter."
• Standing. Where you stand, against OSHA and against the peer network, in four cards: Training Depth (current certifications over total, plus a peer figure when available), Policy Currency (the same percentage as the Explore tile, with how many policies are within cycle), Insurance (active policy count, plus your EMR color-coded, red at 1.25 and above, amber above 1.0, green at or below), and OSHA 300 Log (your recordable count, plus days since the last one). A cert attention list below surfaces anything expired or expiring within 30 days.
• Exposure. Framed as "where capacity is at risk this week, not a list of records, a list of decisions," in three buckets: Cert wall (certifications expiring in 30 days, "book renewal classes now"), Insurance renewals (policies expiring in 60 days, "broker quoting window is open"), and Corrective actions past due (open incidents whose fix date has passed, "closure is the fastest EMR lever"). Each bucket names its own honest empty state when there is nothing to show.
• Benchmarks. A table comparing your numbers against the peer network across six metrics: incident rate, days since last recordable, near-miss ratio, training compliance, corrective-action closure days, and policy review currency. Each row shows your value, the peer value, and the delta, colored to reflect whether the direction is good or bad for that particular metric (a lower incident rate is good, a higher training-compliance percentage is good). A footer note is explicit about the limits of peer comparison here: peer figures apply only where a cohort has formed on that metric, absolute OSHA regulatory thresholds are the floor regardless of how peers are doing.

How data flows in

Nothing in Safety is created by clicking a button labeled "add." The Add Data button in the page header opens the same universal capture flow every section uses, upload, photograph, paste, dictate, or forward, tagged to Safety so whatever you send lands as an incident report, a policy document, or a certificate of insurance. Beyond that:

• Forwarding. Set up auto-forwarding so training confirmations and renewal notices route in on their own; see Forwarding documents.
• The four launch decks. Safety, Corrective Actions, Safety Policies, and Emergency Plans each carry their own flat entry form, for reporting an incident, setting a corrective action, or writing a policy or plan directly, rather than through a document upload.
• Connecting your tools. See Connecting your data for the full picture of every inbound path.

Best-practice example

Say your Days Without Recordable reads 41, with an incident rate above your peer figure and one corrective action already 12 days overdue. Open the Exposure tab first, the "Corrective actions past due" bucket names the overdue item directly. Close that loop through the Corrective Actions launch tile, set a real fix and a due date if one is not on file yet, and mark it done once it is finished. Then check Standing: if your EMR reads amber or red, that is a workers' comp renewal conversation worth having before the policy comes up in Insurance Due. The Benchmarks tab tells you whether your incident rate is a genuine outlier against the peer network or within normal range for a business your size, that context is what turns "an incident happened" into "here is what to actually change."

Related reading

• Certifications: activation and Premier gating: the two-gate mechanism (Section Activation, then Premier tier) that applies identically to Safety.
• Certifications overview: where individual and firm credential tracking lives now.
• Compliance overview: the sibling section for regulatory frameworks and audits.
• The decision workspace: how a safety finding becomes a concrete plan you can act on.
• How benchmarks work: how the peer figures on the Benchmarks tab are built.
• Forwarding documents and Connecting your data: every way records land in Verinode.