SIX WEEK PILOT NOW OPEN
Run it on your own site for six weeks
Pay the implementation and nothing else. If it is not right at the end of the six weeks you stop, with no subscription and no notice period.
Per month, per site
To your first live permit
Application, database and backups
Delivered to some of the biggest names.
Your data stays in the EU
Paris and Amsterdam, never replicated out
Encrypted in transit and at rest
TLS everywhere, AES-256 on storage
A person approves every agent action
Every change carries the name of whoever approved it
Seveso and CAD aligned
Designed against European regulation
DPA ready today
A template your legal team can read now
THE PROBLEM
Most of the incidents we’ve read up on didn’t involve careless people. Someone filled in a permit from memory because the risk assessment sat in a different folder, or an isolation was signed off the way it had been signed off last time. The paperwork existed, but it was just spread across three systems that never had to agree with each other.
Source: Health and Safety Executive, cited in J. Gould, Permits-to-Work in the Process Industries, IChemE Symposium Series No. 151.
of accidents in the sector occur while maintenance work is underway.
Concurrent work is invisible
Two crews, one shared hazard, and no system that notices they have met.
Hazard data never reaches the job
The risk assessment is complete, correct, and sitting in a folder. The permit gets filled in from memory.
Isolation depends on who is on shift
The plan is only as reliable as the technician who has done it before, and the system takes their word for it.
The pattern
The same things go wrong in very different places.
A refinery and a loading dock have almost nothing in common, and yet the same complaints come up on both, which suggests these aren’t really industry problems at all.
Records that do not talk to each other
The risk assessment lives in one place, the permit in another, the isolation plan in a third. Each one is correct on its own, and they can drift apart without anybody noticing.
Contractors on your site, outside your system
The people doing your highest-risk work often have the least access to your procedures. They arrive, they work, and your records catch up afterwards if they catch up at all.
Audit preparation from a standing start
Weeks spent assembling evidence that already exists, because it exists in six places and none of them are joined up. The work was done properly. Proving it is the expensive part.
Three tools, three records, and no single place where they have to line up, until now.
THE GAP
The Missing Middle of Industrial Safety.
ONE END OF THE MARKET
Budget tools
Cheap entry price. Rarely built by practitioners. The add-ons arrive once you are committed.
YOU ARE HERE
European industrial SMEs
Same regulatory obligations. A fraction of the budget. Nothing built for the size you actually are.
THE OTHER END
Enterprise platforms
Real capability, priced and architected for Tier 1 multinationals with dozens of sites.
THE PLATFORM
Permit to Work
An ePTW workflow that routes quickly without letting anything through unchecked. Hazards come in from the register rather than being copied off an old permit, and every signature is time-stamped against the job.
- HSG250 and Seveso aligned authorisations
- Hazards arrive from the register automatically
- Site verification signed before activation
- Live view of every permit open on site
A DAY IN THE LIFE
Hazard identified
Already in the register with its controls and PPE attached.
Permit raised
Hazards populate from the register with their controls attached.
Isolation planned
Energy sources mapped, sequence generated from the template.
Locks verified
Each lock signed for as it goes on, in order.
Permit opens
Only now. Work proceeds, visible to the whole site.
Closed and provable
Full trail, retrievable in seconds when an inspector asks.
The hours you get back, priced at your own rate.
The headline counts administration only, because that is the part you can measure. Incident and insurance value is shown underneath, separately, and left out of the return on purpose.
Your site
Set these to match how it actually runs
Where the hours come from
Valued at 60% of €50 an hour
| Workflow | Hours | Value |
|---|---|---|
| Permits | 208 | €6,240 |
| Isolations | 83 | €2,496 |
| Risk reviews | 60 | €1,800 |
| Audit preparation | 16 | €480 |
| Action follow-up | 364 | €10,920 |
| Total | 731 | €21,936 |
Industrial SME
Confidence: HighEvery input sits inside its typical range
731 hours released a year. We value 60% of them, at €50 an hour, which is €21,936. The licence is €12,000 a year for this segment, one site.
One annual price per site
No per-user, per-permit or storage fees. Anything outside the standard plan is priced in your quote.
Included
- Permits to work, isolation records, risk register, actions and audit dossiers
- Hosting, storage, backups and notifications
- Standard support and every product update
- Unlimited records within the volume range for your segment
- 150 editors, unlimited read-only users
Quoted separately, upfront
- Implementation and onboarding, from €3,000, charged once
- Agentic AI add-on, priced per site
- Integrations: CMMS, ERP, DCS, SSO
- Data migration from paper, Excel or another system
- Multi-site, cross-site SIMOPS, custom approval matrices
- Validation, dedicated environments, enhanced SLA
Anything outside the standard plan is listed and priced in your quote, so the number you agree is the number you pay.
Download as a PDF.
Every line, every figure you entered, and the sources behind them.
This calculation is based on your own input and is used for estimate purposes only.How this is calculated
The headline counts administration only. Incident, insurance and downtime value sits in its own box and never enters the return, because it’s the part a finance director is right to discount. Fines and enforcement avoidance are excluded altogether, as is any suggestion that software prevents an incident on its own.
Three of these numbers are ours: the minutes saved per record, the 60% cash-realisation factor, and the volume caps with the five-times value ceiling. We'll swap them for real data after the pilots. They don't come from a standards body.
Hazard regime changes the routing and nothing else. A Seveso site doesn't get a multiplier here, because no published figure supports one. Our software reduces administration rather than removing it, can't verify that an isolation was physically correct, and doesn't make your organisation compliant with anything.
Sources: HSE HSG250, permit to work. AIChE CCPS process safety metrics. Directive 2012/18/EU, Seveso III. Eurostat ESAW 2023. EU-OSHA, cost of work-related accidents. Eurostat 2025 hourly labour costs. These frameworks support the approach. None of them endorses AdvantageEHS or publishes the figures above.
PREDICTABLE PRICING
Run it on your own site for six weeks
Pay the implementation and nothing else. If it is not right at the end of the six weeks you stop, with no subscription and no notice period.
One-off implementation
Depending on site complexity, and quoted in writing before you commit. This is also what a six week pilot costs.
Subscription, per site
Billed annually. All three core modules included. Twelve month minimum once you continue after the pilot. Agentic AI is priced separately.
✓
All three core modules
✓
Unlimited read-only users
✓
Mobile browser access
✓
Fast and reliable support
✓
All core module updates
✓
Onboarding and configuration
AGENTIC AI
BETA
Agentic AI That Asks Before Acting
Data Harmoniser
Spots the same asset entered three different ways and proposes one clean record.
Tag Interceptor
Recognises field language and offers the official equipment tag before a duplicate exists.
Risk Propagator
Pushes a newly reported hazard to every other team working that same asset today.
Regulation Watch
Flags when an exposure limit in your matrix falls out of step with current EU rules.
Advisory only
An agent proposes and a person accepts. Every change to a record carries the name of whoever approved it.
Never used for training
Your data is not used to train or tune any model. If that were ever to change for a specific customer, it would require that customer’s explicit consent first.
Inference stays in the EU
Prompts are processed inside the European Union. Nothing about a hazard, a permit or a person is sent outside it to be reasoned over.
Every suggestion goes to a named person for a yes or a no, and that decision is recorded in the audit trail like any other action. Human oversight is the principle the EU AI Act is built on, and it’s how we’ve built this.
Agentic AI sits alongside the three core modules as an add-on, and is priced separately.
FREQUENTLY ASKED QUESTIONS
How long does implementation actually take?
−
Usually days rather than months. Browser only, so there’s no server to provision and no infrastructure request for your IT team.
Do the three modules really work together, or are they separate tools?
−
They are one system. The Risk Register feeds hazards into the permit automatically, and the permit cannot activate until the corresponding isolation has been verified step by step. That interlock is the reason the product exists.
Where is our data stored?
+
Does the AI ever act on its own?
+
Our works council is cautious about monitoring software.
+
Early adopter pricing is open to the first five sites. Four places remaining.