Home · Solutions · Legal & compliance · Calibrations and inspections chased before they expire
Solution · Legal & complianceA dispenser without a valid calibration does not surprise the inspector, because it would have surprised the system first
Calibrations and inspections chased before they expire
A robot keeps each station’s equipment register with the validity dates of every regime, opens service cases in advance, chases the protocols after the work and maintains one view: what is valid, what is expiring, what is overdue.
Executive summary
Every station is a collection of equipment under several control regimes at once, with deadlines on different cycles and paperwork in binders. One overdue document can stop sales.
A central register with the validity dates of all regimes, service cases opened in advance and protocols completed automatically after the work.
A station inspection starts with printing the set from the system, not with searching binders. Escalations run up the responsibility lines before a deadline passes.
the equipment register with deadlines; contractors’ inspection calendars; protocols in a DMS; cases and the compliance view in Microsoft Teams
Business problem
Fourteen calendars per station
From a compliance standpoint a petrol station is a collection of equipment, each piece living in its own regime: fuel dispensers with periodic calibration, tanks and installations under technical inspection, oil separators, vapour recovery systems, electrical and lightning installations with reviews, extinguishers and generators. Cycles run monthly, yearly and multi-year, and different contractors run them.
In practice the deadlines are managed by station managers in notebooks and by the technical director in a spreadsheet that is current on the day somebody last updated it. The documentation, protocols, decisions, certificates, sits in binders on site, in versions nobody sees centrally.
The consequences run two ways. A missed deadline means, at best, a disabled dispenser and takings lost until recalibration; worse, a fine and post-inspection orders; worst, an environmental event on a device that formally had no valid review. And even when deadlines are kept, an inspection can turn into a hunt for a paper that is certainly somewhere.
Nothing in this process is intellectually hard and everything is organisationally hard: hundreds of dates, dozens of contractors, documents scattered. Which is exactly the profile of work where automation is dependable and a human, by nature, is not.
How it works today
Below is what the work looks like before anything is automated.
- PersonThe station manager keeps his own deadline notebook; head office has a spreadsheet updated occasionally
- Risk of errorCycles of different regimes pass independently; no single person sees them all
- PersonA calibration or inspection order goes out when someone remembers, or when a letter arrives
- WaitingAfter the work, the protocol travels to a binder at the station; head office sees the invoice, not the document
- Risk of errorAn inspector asks for a document nobody can find, although the work was done
- Risk of errorAn overdue item surfaces at an inspection or a failure, that is, at the most expensive possible moment
Why the current process costs more than it appears
The bill that never shows up in a budget.
- A disabled dispenser waits days for recalibration: that nozzle’s takings simply vanish, and customers fill up next door, at the competition.
- Services ordered last-minute are pricier and scheduled worse; the contractor knows you have no choice.
- The technical director’s and managers’ hours spent watching dates and hunting papers, in a rhythm that still gives no certainty.
- Every inspection is a documentation lottery even though the work was done and paid for; you pay a second time, in stress and orders.
Cost of inaction
The first two rows are countable and usually enough for the project to defend itself. The third is of another category: at an environmental event or an accident, the lack of a valid review changes the company’s position towards the insurer and the authorities in a way no page prices.
There is also the opposite cost, rarely counted: services performed too early, “to be on the safe side”. A register with true dates lets you plan close to the line with a safe margin instead of paying for overzealousness.
A model organisation with realistic proportions – the numbers exist so you can run the same maths on your own data; they are not a client result.
A network of forty owned stations: a dozen or more devices under regimes per station, a dozen or more contractors for calibrations, inspections and reviews, documentation in binders, Microsoft 365 at head office.
214 validity deadlines a year, around 600 documents (protocols, certificates, decisions) to complete and store.
A central spreadsheet plus managers’ notebooks; reactive service orders; documents at the stations; a two-day mobilisation before every announced inspection.
Two positions disabled last year after missed deadlines; three inspections with documentation orders despite the work having been done.
The robot keeps a central equipment register with the dates of all regimes, opens service cases with regime-appropriate lead time, orders the service from the assigned contractor after approval, chases the protocol after the work (reads it and updates the next deadline) and maintains a compliance view per station.
In the modelled case the word “overdue” leaves the operational vocabulary, and an inspection starts with printing the set. Model numbers, not anyone’s ledgers.
Proposed solution
The foundation is the register, not the robot: we inventory each station’s equipment, its regimes, cycles, contractors and required documents. It is a one-off stocktake done together with your technical department, and the register is versioned from day one.
From then on the robot watches the dates: with regime-appropriate lead time it opens a service case, prepares the order to the assigned contractor and sends it after a person’s approval. After the work it chases the document: the protocol arrives by email or is scanned at the station, the robot reads it, files it under the device in the DMS and updates the next deadline. A protocol missing beyond a reasonable time escalates to the contractor, instead of being discovered at an inspection.
The compliance view shows each station on one card: devices, validities, open cases, documents. An announced inspection means printing the set; an unannounced one, exactly the same. Escalations run up the lines: manager, region, technical director, with reaction times recorded in the case.
UiPath Orchestrator: per-regime schedules, case queues, retries and an audit trail; UiPath Document Understanding for reading protocols; UiPath Integration Service connectors for Microsoft Teams and Outlook 365; UiPath Action Center cases
The stocktake and the versioned equipment register, per-regime lead-time logic, service cases with orders, protocol reading with deadline updates, the device DMS and the per-station compliance view
The contractors’ mailbox and station scans; SharePoint/DMS for documentation; optionally the ERP for service orders and invoices; inspection calendars
How the automated process works
- SystemThe register knows the dates of all regimes; an approaching deadline opens a case with the right lead time
- AutomationThe robot prepares the service order to the assigned contractor with the device’s data
- PersonA person approves the order; atypical cases (replacement instead of calibration) are decided by the technical department
- AutomationAfter the work the robot receives the protocol, reads it, files it under the device and updates the next deadline
- SystemA protocol missing on time escalates to the contractor; no reaction goes up the line
- PersonThe per-station compliance view serves inspections, audits and service budget planning
Human-in-the-loop model
Automation handles
- Watching all the dates and opening cases with regime-appropriate lead time
- Preparing orders, receiving and reading protocols, and updating deadlines
- Escalations, documentation completeness and the per-station compliance view
People decide
- The equipment and regime register and every change to it (new device, replacement, retirement)
- Order approvals and atypical decisions: replacement, modernisation, contractor change
- Relations with contractors and authorities, and conclusions from post-inspection orders
Before and after
Systems and integrations
The stack is short on purpose: one engine, one execution layer, one place where a person decides.
Inputs
- the equipment register with regimes and cycles
- contractors’ correspondence and protocols
- document scans from the stations
- inspection calendars
Automation layer
- UiPath Orchestrator
- UiPath Robots
- UiPath Integration Service
- UiPath Action Center
Target systems
- service cases and post-approval orders
- the device DMS with complete documentation
- the per-station compliance view in Microsoft Teams
Human touchpoints: UiPath Action Center cases in Microsoft Teams; the compliance view as a Teams tab; a quarterly register review with the technical department
Technologies used
per-regime schedules, case queues, retries, audit
Areading protocols and certificates from PDFs and scans: dates, device numbers, results
Aorders to contractors, protocol receipt from email, escalations
Aorder approvals and atypical decisions where the technical department already works
Aper-device documentation with metadata and versions, in your Microsoft 365 tenant
AHow we deliver this
From this description to a running process
We deliver this solution through our process automation service: first we measure the process on company data, then we build, test and maintain the robots.
Industries
Case studies
Delivered client work is described under Case studies.
Illustrative economic model
Numbers you can check against your own data.
The disabled-positions row assumes that misses happen and that they stop: both assumptions should be checked against your own history. The value the model leaves out is the company’s position towards the insurer and the authorities on the day of an event, when a complete set of valid reviews is worth more than this whole arithmetic.
Run the numbers on your data
An illustrative estimate based on your inputs. It models freed capacity, not promised savings.
Business benefits
- The word “overdue” leaves the operational vocabulary, because cases open before the deadline
- Services ordered in advance are cheaper and better scheduled
- Protocols are where the inspector looks for them: filed under the device, with the next deadline set
- Positions disabled after misses stop eating the takings
- The service budget is planned from the register, not from surprises
The management view
- Infrastructure compliance stops depending on the notebooks and memory of forty managers
- Every device has a history: services, documents, decisions, escalations
- The company’s position on the day of an inspection or an event is built daily, automatically
Board-level KPIs
devices with valid deadlines (share) · cases opened with the right lead time · protocols completed on time · positions disabled after misses · time to prepare a station for an inspection
Security and governance
The automation has exactly the permissions it needs. Not one more.
- The robot changes nothing in the devices or station systems; it keeps the register, the cases and the documentation
- Every deadline change has a source: the protocol it was read from, with a version and a date
- The equipment register is versioned; structural changes require the technical department
- Personal data is limited to signatures on protocols; cases deal in devices and stations
- Documentation and cases stay in your Microsoft 365 tenant, and the robots run in the EU region of UiPath Automation Cloud
Why now
Control regimes multiply rather than shrink; every new duty is another calendar glued onto a spreadsheet that already does not close today.
Station margins do not forgive disabled positions: a single miss can consume the yearly cost of the whole watching system.
The documents already exist: protocols, certificates, decisions. All that is missing is the process that reads them, files them and extracts the next deadline.
Relevant executive roles
Instead of watching dates, they manage exceptions and the budget; the register watches itself
The per-station compliance view exists at any moment, not after a two-day mobilisation
Positions disabled after misses vanish from the reports, and inspections stop being a lottery
Common questions and objections
Some do, within their own scope and in their own interest. But the responsibility is yours, the regimes are several, the contractors a dozen, and none of them watches the whole. The register does not replace the contractors; it removes your reliance on every one of them working.
It is the biggest item of the implementation and we say so openly. It is done once, methodically, station by station, partly from documents that already exist (service invoices, old protocols). The alternative is never doing it and watching deadlines for devices whose list does not exist.
If the CMMS has complete data and the regimes’ dates, it stays the source: the robot reads from it and adds document chasing, escalations and the compliance view. Usually though a CMMS knows about technical maintenance, not calibrations and statutory inspections; then we complete it rather than duplicate it.
When this is not the right solution
- A few stations and one trusted do-it-all contractor: a shared calendar will do
- Technical service fully outsourced with contractual penalties for misses: then you watch the contract, not the deadlines
- No readiness for the stocktake: without a reliable register at the start, the automation would watch a fiction
A question for the next management meeting
How many devices in our network have an expired deadline of any regime today, and how long would it take us to establish that number?
Implementation approach
Scope without ambiguity, before anything is signed.
We deliver
- The equipment stocktake and a versioned register with all regimes’ dates
- Service cases with per-regime lead times and post-approval orders
- Protocol reading with deadline updates and the device DMS
- The per-station compliance view and escalations up the responsibility lines
- Post-go-live care and a quarterly register review through the first year
We need from you
- Access to the current spreadsheets, service invoices and binders for the stocktake
- A designated owner of the register in the technical department
- The contractor list with scopes and the mailbox for service correspondence
Stages
Discovery
A pilot stocktake of three stations: devices, regimes, document state, the real cycles
Rules
Per-regime lead times, escalation lines and order templates approved by the technical department
Build
The register, the cases, protocol reading, the DMS and the compliance view
Parallel run
The stocktake of the rest of the network while the system runs on the pilot
Go-live
The whole network in the register; spreadsheets and notebooks become history
A quick win on the pilot, full value after the network stocktake. Effort depends mostly on the state of today’s documentation; the automation layer itself is simple.
The inspector asks for dispenser six’s calibration certificate. It is in a binder. One of them. At one of the stations.
Send us your current deadline spreadsheet and one station’s documents. We return that station’s register in our structure, the deadlines needing action within a quarter, and what is missing from the documents, before somebody else asks.
Send one station’s register for a reviewThe neighbouring process usually has the same problem
Driver and vehicle documents expire in fourteen calendars. This page does the same for people and vehicles.
See the solution Operations & productionTanks reconciled nightly, losses visible by morningThe gauges say one thing, the tills another. A daily balance says why.
See the solution Legal & complianceFuel licence reports ready before the deadlineThe twentieth of the month stops being panic day for the licensing desk.
See the solution