ZedNerve Platform
See what a change or delay does to the outcome and who holds each step. AI agents check and recommend under agreed rules; authorised people decide. Execution is coordinated from one record.
Your systems stay where they are.
Capabilities
Illustrative example. Fictional data.
See where execution is slipping, who holds the work and what the delay costs.
See it in the example : Nerve CenterExplore changes to who does each step, and their projected impact.
See it in the example : AI Operating ModelFollow required inputs, reviews, authorised decisions and the next steps, one by one.
See it in the example : JourneyEvaluate how people, AI agents and their supervisors handle reviews.
| Reviewer | On time | Average hold (days) |
|---|---|---|
| Readiness Agent AI agent | 99% | 0.1 |
| Analyst 1 Person | 88% | 1.2 |
| Analyst 2 Person | 76% | 1.5 |
| Analyst 3 Person | 63% | 2.1 |
Supervisors are measured on how they supervise AI
Procurement lead Authorised person
A view of how reviewers work, not of programme outcomes. Holds are per supplier evidence pack.
Improve the operating model
The same record shows which steps recur. The example simulates one change to a recurring check, across 4 rollout waves.
Authority
Every action in ZedNerve is one of four kinds, and each kind has its own rule for who may take it.
Only authorised people approve, reject or return. Financial approval limits your organisation sets decide which person signs, for example a manager up to an agreed amount and a director above it; agents never decide.
Allowed under an approved policy, such as a reminder when work passes its agreed time limit.
Prepared by AI agents or people, such as a readiness check. They inform a decision; they never make it.
Approvals, rejections and returns, made only by an authorised person. The amount involved decides which person that is, never whether an agent may act.
An overdue approval can pass to the named backup approver under the agreed rules. An escalation to the responsible manager is drafted by an agent and sent only once a person confirms it.
Each process runs through gates with named roles. Every action is checked against the person's authority at the moment it happens, and an attempt outside a role is refused and recorded.
Watchers detect a missed time limit, or work that went ahead past a gate that was never approved. Reminders go out automatically under the agreed rules, and an overdue approval can pass to the named backup approver. An escalation reaches the responsible manager only once a person confirms it.
Every action and every refused attempt is recorded with the authenticated identity of who acted and when. Decision records are append-only: entries are never edited or deleted, and they are hash-chained, so any later change shows.
Capability status
The ZedNerve Platform is in pre-production. It works alongside your finance, HR and delivery systems rather than replacing them, and which systems connect is agreed per organisation. This is where each part stands today.
| Part | Status | What it means |
|---|---|---|
| The four screens | Shown in the demo | Nerve Center, AI Operating Model, Journey and Reviewer Performance, with fictional data, in a conversation. The demo currently follows a property-development programme. |
| Arabic and English product screens | Built, pre-production | Product screens in both languages, right to left included. The demo shown in conversations is in English. |
| Audit history | Built, pre-production Planned | Who acted and when, refused attempts included, hash-chained so any change shows. Export is planned. |
| Reminders and escalation | Built, pre-production | Reminders under agreed rules; overdue approvals can pass to a named backup; escalations are sent once a person confirms. |
| Forecasts | Built, pre-production Simulated | Forecast dates for pending steps from your own history, with the basis shown. The demo's operating-model projections are simulated. |
| Assessing a change | Built, pre-production Planned | Advisory today: we assess a change, what depends on it and the possible responses with you in the engagement. Forecast dates from your own history are built; comparing response options side by side in the platform is planned. |
| PDF and Excel exports | Built, pre-production Planned | Built for the diagnostic report. An Arabic PDF and further exports are planned. |
| Integrations | Built, pre-production Planned | A connector framework and history import from CSV or Excel. Live connectors are planned, scoped per organisation; none is live in the demo. |
| Required inputs before approval | Built, pre-production | Rules block an approval when required inputs are missing. |
| AI document checks | Simulated Planned | In the demo, the readiness check's findings are replayed; no AI model runs. In the product, AI checks are planned. |
| Outcome measurement | Configured per engagement | Baseline, targets and the rules that price delay agreed with you; actual results measured and reported through the engagement. |
| Data on this site and in the demo | Simulated | Fictional projects, people, AI agents and figures. |
Hosting, data handling and integration requirements are agreed for your organisation and operating environment.
Measurement
We agree each measure with you at the start. It keeps four values, and they stay distinct: a forecast is never shown as a result.
A pending step gets a forecast completion date from your completed steps of the same kind, and the forecast shows what it is based on.
We price delay with rules agreed with your organisation, for example an agreed annual saving ÷ 52 for each week, and show the arithmetic. No hidden model is involved.
After go-live, we compare each measure with its baseline, its target and its forecast, measured the same way as the baseline.
Finance, audit and leadership read status, decisions and measures from the same record. The diagnostic report exports to PDF and Excel; data and audit exports are planned.
Wave 1 measures
| Measure | Baseline | Target | Forecast | Actual |
|---|---|---|---|---|
| Turnaround, working days per request | 6.0 | 3.0 | 3.5 | Not yet measured |
| Requests within service level (5 working days) | 64% | 90% | 85% | Not yet measured |
| Requests resolved right first time | 82% | 95% | 92% | Not yet measured |
Actual results come only from real operation, measured 8 weeks after each go-live the same way as the baseline.
FAQ
No. AI agents check submissions, prepare drafts and recommendations, and send routine reminders under agreed rules. Only authorised people approve, reject or return, and financial approval limits your organisation sets decide which person signs. An attempt by an agent to act outside its role is refused and recorded.
Hosting, data handling and integrations are agreed for your organisation and operating environment. Forecasts run on your own delivery history.
No. The platform supports every ZedNerve engagement and can also be piloted on its own.
A pilot runs one process for about 12 weeks, in an environment agreed with you. Its gates, roles and cost-of-delay rules are then kept as a template for the next process of the same kind.
The systems you already use stay in place. History can be imported from CSV or Excel, and live connectors are scoped and agreed per organisation. In the demo, connectors are shown as planned or partner-provided, and none is live.