"Signals: how Verinode watches your data (the Pipeline section)"

A Signal is a single, machine-detected observation about your business: a job that has gone quiet, a carrier balance aging past that carrier's own average, a certification about to lapse, a vendor…

9 min read·Updated July 13, 2026
On this page

What a Signal is

A Signal is a single, machine-detected observation about your business: a job that has gone quiet, a carrier balance aging past that carrier's own average, a certification about to lapse, a vendor whose spend has drifted above what the peer network pays for the same work. Every one of these lives as a row Verinode writes to your operator record the moment a detector notices it, and every one carries a domain (Vendors, Equipment, Carriers, Clients, Billing, Team, Reputation, Field, Margin, Compliance, Certifications, Safety, Recruiting, Processes, and a few narrower ones like Weather and Competitive), a severity (Critical, Warning, or Info), and a status that tracks whether you have seen it, acted on it, parked it, or it has resolved on its own.

Verinode does not create these situations and it does not decide what you do about them. It reads the jobs, invoices, documents, and connected data you already have, watches for the patterns a fractional COO would flag, and surfaces them as decisions. You take the action, or you do not.

This article explains where this used to live, where it lives now, and how the detection underneath it actually runs. For how to work the resulting decisions day to day, see the Feed and the decision workspace.

Why /pipeline and /signals redirect now

Early in the product this was a section called Pipeline, then Signals, with its own page in the sidebar. Both names described the same idea: a flat list of everything the detection engine had found. As the product grew, that flat list split into two purpose-built surfaces that do the job better:

  • Feed (/feed) is the daily read: a scrollable deck that mixes your decisions in with industry articles, vendor news, and weather alerts, so you get a rhythm of what changed today.
  • Decisions (/decisions) is the full gallery: every open, parked, acted-on, and resolved decision in one filterable, sortable list, with running totals of your impact.

Because of that split, the old routes are kept alive only as redirects, so nothing that ever bookmarked or linked to them breaks:

  • /pipeline redirects straight to /signals.
  • /signals then redirects again. If the link carried a specific decision (an id in the URL, the shape used by deep links from emails, the daily digest, or an in-app "Act" button), it sends you to that exact decision at /decisions?id=<id>. Without an id, it sends you to the Feed with the Decisions filter pill already selected, at /feed?filter=decisions.

Note

There is no page left at /pipeline or /signals to look at directly, both are pass-through redirects with no UI of their own. If you have an old bookmark, a browser favorite, or a saved link in an email to either address, it will still get you to the right place, just one hop further than before.

The sidebar itself reflects this: there is no "Pipeline" or "Signals" entry anymore, only Feed and Decisions, sitting at the top of the nav above Network Playbooks and the rest of the sections.

Where to find it today

Open Feed from the sidebar (/feed) for the daily mixed read, or Decisions (/decisions) for the full log. Both draw from the same underlying Signals.

The Decisions filter, inside Feed

At the top of the Feed there are four filter pills: All, Decisions, Content, Events. Clicking Decisions narrows the deck to only your decision cards (dropping the vendor news, articles, videos, podcasts, and weather alerts). This is exactly what landing on /feed?filter=decisions from an old Signals link does automatically. Beside the filter pills, a separate time-range control (Today, This week, 30 days, All time, default 30 days) fences how far back the deck reaches, so an active operator's history does not turn into an endless backlog.

When there is nothing new, the Feed shows: "All caught up. Verinode IQ is continuously analyzing your data and scanning industry sources. New decisions, insights, and updates will appear here as they surface. Check back soon. Your next briefing is building." If a filter has no matching cards, it reads simply "No items match this filter."

The Decisions workspace, in full

/decisions is the gallery view of every signal-driven decision you are tracking, resolved ones included. It is built for triage rather than a daily scroll: a domain dropdown (the same humanized list: Vendors, Equipment, Carriers, Clients, Billing, Team, Reputation, Field, Margin, Compliance, Certifications, Safety, Recruiting, Processes), a status filter (All, Pending, Acted, Parked, Ignored, Resolved), and a sort control, sit above the list. A stat strip above the list totals what is happening across your whole log: Pending (need your attention), Acted (decisions you have taken), Parked (saved for later), Resolved (closed out), and Impact addressed, the estimated annual dollar value of what you have resolved. Full mechanics of that workspace, including how orchestrated decisions differ from single-signal cards, are covered in the decision workspace and acting on decisions.

Tip

If you arrived here from an email digest or an "Act" button elsewhere in the product and the link carried a decision id, you already skipped the gallery and landed straight on that one decision. Use the back arrow or the Decisions sidebar entry to get back to the full list.

How detection actually runs

Signals are not something you request or refresh on demand. A scheduled pass runs the full set of detectors for every operator, working through your data domain by domain. Each detector is independent and non-fatal: if one detector hits an error, it is logged and skipped, and every other detector still runs, so a single failure never blocks the rest of your Signals from updating.

Across that pass, the detectors cover the same domains you see in the Decisions filter, among them:

  • Vendors: renewal windows coming up, spend drifting above what similar operators pay for the same service, coverage gaps, a vendor blocked or required by a carrier program you belong to.
  • Carriers: SLA misses, denial and pushback patterns by reason and by carrier, adjuster-level patterns (approval rate, response time, escalation outcomes) once enough of your supplement history carries an adjuster.
  • Clients: concentration risk, stance pressure.
  • Team: single points of failure, certification lapses stacking up, bench depth, recruiting pipeline health.
  • Certifications: expiring credentials, program lockouts, gaps against what your peer cohort typically holds.
  • Compliance: audit exposure, findings still open, regulation adoption lag.
  • Safety: incident rate, EMR trend, insurance and certification gaps.
  • Equipment, Fleet, Facilities: aging assets, calibration and inspection due dates, lease and renewal windows.
  • Margin: labor burden drift, crew utilization drift, job types lagging your own book or the peer network, correlations between how long a job's stages take and its margin or satisfaction outcome.
  • Processes: service lines missing a documented procedure, a procedure that diverges from what peers typically do, stages drifting slower against your own recent history, jobs stalled with no trace of activity, receivables billed with no follow-up on record.
  • Reputation: review score and response-time trends, unclaimed profiles, negative topic clusters.
  • Weather and Competitive: restoration-relevant weather risk against thin coverage, and gaps against research or the peer network more broadly.

For every operator on every pass, this rule-based layer runs in full, including on the free tier. On top of it, once you are past the free tier, a second reasoning layer runs: it clusters related atomic signals (several team gaps that add up to one hiring decision, several adjuster patterns that add up to one escalate-or-accept call) into a single weighed decision, rewrites the card into a plain-language title and consequence line, and runs a feasibility check that clamps any dollar estimate to a realistic share of your business and can suppress a decision entirely if its premise does not hold up. That is the difference between a raw detector output and the polished card you see on Decisions: the underlying observation is the same, the framing and the sanity-check are the added layer.

Tip

A signal is not necessarily new just because it appears in Feed today. Verinode tracks when a pattern was first detected separately from when it was last seen, and updates the same signal in place as long as the condition persists, rather than firing a fresh duplicate every night. It disappears from your open list on its own once the underlying condition clears (the job resumes activity, the certification renews, the metric recovers), or when you dismiss it.

Evidence behind a signal

Many signals carry supporting evidence beyond the headline: an observed value from your own data, how that compares to a published research benchmark, and how it compares to the anonymized peer network where enough contributors exist to compare against safely. Where a peer comparison is not yet statistically safe to show, Verinode holds it back rather than exposing a comparison built on too few operators. See benchmarks overview, how benchmarks work, and reading a benchmark for how those comparisons are built and gated across the platform.

Some signals also carry a confidence read and, once you resolve one, an outcome: what happened, and where Verinode can put a dollar figure on it, that verified figure is what rolls up into the Impact addressed total on the Decisions page.

Empty states

  • Feed, nothing new: "All caught up. Verinode IQ is continuously analyzing your data and scanning industry sources. New decisions, insights, and updates will appear here as they surface. Check back soon. Your next briefing is building."
  • Feed, filter has no matches: "No items match this filter."
  • Decisions, nothing in a domain or status filter: the list simply comes back empty; adjust the domain or status dropdown, or check All to see everything.

None of these are broken screens. An empty Signals surface at a new account, or right after you clear a backlog, means the detectors have nothing to flag yet, not that something has failed.

Best-practice example

You get an old email digest link that points at /signals?id=…. Clicking it takes you straight to that one decision inside /decisions, no detour through a list. The next day, you open Feed out of habit, tap the Decisions pill (the same place a bare /signals link would have sent you), and work through what is new since yesterday inside the 30-day window. Once a week, you switch to the full /decisions gallery, filter by status Pending, sort by newest first, and clear the backlog domain by domain, starting with anything marked Critical.

  1. 1Bookmark Feed (/feed) for your daily read and Decisions (/decisions) for the full log; retire any old /pipeline or /signals bookmarks, they still work but add an extra hop.
  2. 2On Feed, use the Decisions filter pill when you want signals only, and the time-range control to widen or narrow the window.
  3. 3On Decisions, use the domain dropdown and status filter to triage: Pending first, Critical severity first.
  4. 4Read the evidence on a card, your own value, the research benchmark, the peer comparison where one is shown, before you act.
  5. 5Act, park, or dismiss. Resolved decisions with a verified dollar outcome roll into your Impact addressed total.

Data sources

  1. 1.pii.signals (detector output, status, severity, evidence). Your business.
  2. 2.intelligence research benchmarks and the anonymized peer network. Verinode reference data.
Was this helpful?