Exception surfacing
Reads activity signals so nobody has to watch a feed, and raises only what departs from a team’s own established pattern.
Zepth Node · Workforce
Workforce Monitoring
Monitoring built on consent, not surveillance — opt-in, clocked-in only, and revocable at any time.
Last updated
Zepth Node module
Workforce Monitoring
AI agent built into the module
Desktop activity agentConsent-first, alwaysActivity alertsRetention controls
Overview
Most monitoring software is sold on the assumption that the employee must not know how it works. That assumption is what makes it corrosive: a workforce that discovers the rules after the fact stops trusting every other system it is asked to use, including the ones that pay it.
Node inverts it. Monitoring runs only with the employee’s explicit consent, only while they are clocked in, and stops the moment either condition ends. The disclosure and the consent both sit on the record, so what was agreed is checkable by both sides.
Consent is a product behaviour, not a policy
A consent that lives in a handbook is a claim. A consent that gates whether the agent runs at all is a mechanism. The difference matters because only one of them still holds when somebody changes their mind on a Tuesday afternoon.
In Node, withdrawal stops capture immediately. There is no administrative step between the employee’s decision and the software’s behaviour, and both the original disclosure and the withdrawal remain on the record.
Opt-in. Nothing is captured until the employee has seen the disclosure and given explicit consent.
Clocked-in only. Capture is bounded by the working session. Clocking out ends it — not pauses it.
Revocable. Consent can be withdrawn at any time, and capture stops when it is.
On the record. What was disclosed, what was agreed and when it changed are all auditable by both sides.
Retention is where monitoring goes wrong
The failure mode is rarely the capture. It is the archive: months of screenshots nobody set an expiry on, sitting in a bucket, useful to no one and available to anybody who gains access. The organisation carries the risk long after the question that justified collection was answered.
Automatic purging on a schedule you set removes the archive as a category. What is kept is bounded by policy the system enforces, rather than by an intention somebody documented.
What this is for, and what it is not
Monitoring answers questions about how work happens — where hybrid teams lose time to tooling, whether a process change helped, whether a workload is sustainable. Those are aggregate questions, and they are answerable without watching an individual.
It is not built to catch people, and it should not be used that way. A workforce that believes it is being watched behaves like one, and the data stops describing the work it was collected to describe.
“Node monitoring is opt-in, clocked-in only, and revocable. It is a productivity and assurance tool, governed by consent and retention limits — not covert surveillance.”
The value
Why it matters
An honest picture of remote and hybrid work, gathered with the employee’s agreement.
Capture bounded by the clock — nothing is recorded outside a working session.
Consent that can be withdrawn, with the withdrawal taking effect immediately.
Retention limits enforced automatically, so you hold what you need and nothing longer.
Capabilities
What you can do
01
Desktop activity agent
A cross-platform agent captures active and idle heartbeats and periodic screenshots — only while the employee is clocked in, and it stops the moment they clock out.
02
Consent-first, always
Monitoring runs only on explicit consent and stops immediately when consent is withdrawn. The disclosure shown and the consent given are both recorded.
03
Activity alerts
Surface the exceptions worth a look instead of asking somebody to watch a live feed — which is both ineffective and the thing everybody fears.
04
Retention controls
Screenshots are purged automatically on the retention schedule you set. Deletion is enforced by the system rather than left to a policy document.
The workflow
How it actually runs
- 1
Configure scope and retention
Decide what is captured, for which teams, and how long it is kept. Retention is enforced by the system, not by a document.
- 2
Disclose and request consent
The employee sees the disclosure in plain language and gives or declines explicit consent. Both are recorded.
- 3
Capture within the session
The agent runs only while the employee is clocked in and stops when they clock out.
- 4
Review alerts, not feeds
Exceptions are surfaced with their reasoning attached, for a human to judge.
- 5
Purge on schedule
Screenshots are deleted automatically when the retention period ends.
AI that does the work
How AI changes Workforce Monitoring management.
Aggregate insight
Answers questions about how work happens at team and period level, which is the level the useful questions live at anyway.
No conclusions about people
The agent draws no inference about an individual’s performance or intent and takes no action. It surfaces; a human decides whether anything follows.
The engineer’s judgment stays in charge; the AI removes the latency and the blind spots.
Best practices
- Write the disclosure in plain language and show it before consent is requested. A consent granted against text nobody read is not consent, whatever the record says.
- Set the shortest retention period that answers your actual question. Screenshots kept indefinitely become a liability long before they become useful.
- Review alerts, not feeds. Watching live is bad practice, bad management and bad use of the data.
- Make withdrawal genuinely available and never attach a consequence to it. A consent that costs something to withdraw is not one.
- Keep monitoring data out of pay decisions unless your policy says so explicitly and the employee has been told.
Dashboards & reporting
Dashboards report aggregate activity patterns by team and period, alert volume and resolution, and consent and retention status across the workforce. Individual-level access follows role-based permissions and is itself logged.
Live dashboards
Drill-down & filters
Export to Excel / PDF
FAQ
Common questions
Can an employee refuse monitoring?
Yes. Monitoring is opt-in and runs only on explicit consent. Consent can also be withdrawn at any time, and capture stops immediately when it is.
Is anything captured outside working hours?
No. Capture is bounded by the clocked-in session. When the employee clocks out, the agent stops — it does not continue in the background.
How long are screenshots kept?
For the retention period you configure, after which they are purged automatically. Deletion is enforced by the system rather than left to a policy document.
Which platforms does the agent run on?
It is a cross-platform desktop agent capturing active and idle heartbeats and periodic screenshots, active only during a clocked-in session.
Does monitoring data affect pay?
Not by default. Pay is calculated from verified attendance, leave and overtime. If your own policy connects the two, that must be stated to employees explicitly.
What does the AI conclude about an individual?
Nothing. It surfaces exceptions against a team’s own pattern with the reasoning attached. Any conclusion about a person is drawn by a human who is accountable for it.
Related modules
Zepth is the construction project delivery platform — it runs construction, procurement and asset management on one record, and does the work: reading the drawings, reviewing the submittals, matching the invoices and flagging the risks, with a human sign-off on anything consequential.
See it on your project.
A short, tailored walkthrough on your real workflow — no generic demo.