Automating Field Service Proof of Work: Parts, Signatures and Evidence That Survives Audit
What auditors and warranty assessors ask for, the Dynamics 365 Field Service records that hold it, a capture flow technicians will use, and an honest method for the gaps.
An auditor or warranty assessor asks for the same things: which asset, which serial numbers, which parts were consumed and fitted, who attended, when, and evidence captured at the time rather than reconstructed later. In Dynamics 365 Field Service that evidence has a home: the customer asset holds the serial, work order products record parts as Estimated until marked Used against the booking, the booking records who attended and when, and photos, service task results and the customer signature sit on the work order. A mobile flow that scans assets and parts and prompts for photos records this without adding minutes per job. Missing evidence can only be reconstructed from surviving sources and labelled as such.
What does an auditor or a warranty claim actually require as proof of work?
Auditors, warranty assessors and regulators phrase their questions differently, but they are asking for the same evidence, and they give much more weight to a record captured during the visit than to one assembled afterwards. If your system cannot answer each row below from a record created at the time, that is the gap to close, and it is a data model and capture problem rather than a reporting one.
| The question | Evidence that answers it | Where it lives in Field Service |
|---|---|---|
| Which equipment was serviced? | A unique asset identity and its location | Customer asset, with functional location where the site structure matters |
| Which serial numbers? | The serial of the unit and of serialised components fitted or removed | Serial number on the customer asset and on the component assets beneath it |
| Which parts were consumed and fitted? | Product, quantity, serial where relevant, and the visit it happened on | Work order products with line status Used, linked to the booking |
| Who attended? | A named, licensed person, not a team name | Bookable resource on the booking |
| When? | Arrival, work and completion times recorded by the system | Booking status changes, timestamps and booking journals |
| What was done and found? | Task results and readings in structured fields | Work order service tasks and incident type results |
| Proof it happened as recorded | Photos, customer signature and an unaltered record | Attachments on the work order, signature column, service report, audit history |
Which Dynamics 365 Field Service records hold the evidence?
The evidence model uses standard Field Service tables. How the installed base, incident types and agreements are designed during a build is covered on our Field Service implementation partner page, so this section concentrates on what each record contributes as evidence and the design choice that decides whether it holds up.
| Record | Evidence it carries | Design choice that matters |
|---|---|---|
| Customer asset | Identity, serial number, parent and child components, account and location | A parent and child hierarchy so a replaced component has its own serial and history; a serial column added where your version does not already carry one |
| Work order | The job, its incident type, the primary asset, the signature and the service report | The primary asset is required, not optional, for work that may be claimed or audited |
| Bookable resource booking | Who attended, and the status timeline of the visit | Every custom booking status mapped to a Field Service status so completion and timestamps behave |
| Work order product | Parts estimated and parts actually Used, with quantity, warehouse and the booking that consumed them | Estimated lines from the incident type, set to Used on the device; never free text in a notes field |
| Work order service task | The checklist and its results | Typed result fields or inspection responses rather than a completed tick box alone |
| Notes and attachments | Photos and documents with the time and the author | Photos attached on the work order, named by step, not left in the device camera roll |
| Audit history | Who changed which value, from what, to what, and when | Auditing enabled on these tables and the columns that matter, with a retention period agreed with compliance |
Why do parts go missing from billing, and how do work order products fix it?
When a customer is billed for the wrong parts, or a warranty claim fails because nobody can show what was fitted, the cause is usually the same: parts were written on paper, or Estimated lines were never changed to Used. There is no separate booking product table in Field Service. A work order product line starts as Estimated when the job is planned and becomes Used when the technician confirms it was consumed, and that line can reference the booking on which it was used. Used lines are what drive inventory adjustments and invoicing; Estimated lines left in place make the job look planned rather than done.
- Pre-populate Estimated lines from the incident type, so the technician confirms rather than searches.
- Mark lines Used on the device during the visit, and link them to the booking so a multi-visit job shows which visit fitted which part.
- Add unplanned parts from the technician van warehouse by scan, so stock and evidence come from the same action.
- Block completion of a booking, or flag it for review, when a job type that always consumes parts has no Used lines.
- Record the removed part as well as the fitted one for serialised components, and update the customer asset so the installed base shows what is in the machine now.
- Reconcile Used lines against van stock and the ERP on a schedule; Solzet integrates with your ERP but does not implement Dynamics 365 Finance, Business Central or any other ERP.
What mobile capture flow records the evidence without adding minutes per job?
Technicians abandon evidence capture that feels like paperwork layered on top of the job, and the result is the gap the auditor finds. The design rule is that evidence is a side effect of doing the work: every status change, scan and confirmation the technician already performs writes a piece of evidence, and the only new actions are a small number of required photos and the signature. The flow below is the one we build to on the Field Service mobile app or a canvas app on Dataverse. If technicians work with poor signal, pair it with the durable offline design in our guide to offline field apps that do not lose data.
- 1. Change the booking status on the device. The technician sets Traveling, In Progress and Completed in the app. Each change timestamps the booking, so who attended and when is recorded as a side effect of work the technician already does, not as a separate form.
- 2. Scan the asset to open the job. A QR code or barcode on the machine opens the customer asset and confirms the technician is at the right equipment. A mismatch between the scanned asset and the asset on the work order is flagged immediately rather than discovered at audit.
- 3. Confirm or correct the serial number. The serial on the asset record is shown for confirmation with one tap, or scanned from the rating plate when it differs. A corrected serial is written back to the customer asset, so the installed base stays true for the next visit.
- 4. Work through the service tasks. Tasks come from the incident type, so the checklist is already on the device. Results and measurements go into typed fields rather than a notes box, which makes them evidence rather than prose.
- 5. Mark parts Used by scanning, not typing. Parts expected for the job are already on the work order as Estimated lines from the incident type. The technician scans what was actually fitted to mark it Used against the booking, adds anything unplanned from van stock, and scans the serial of serialised parts, including the part removed.
- 6. Take photos at the points the process defines. The app asks for a photo only where evidence is required, such as the rating plate, the fault before work and the completed installation, and attaches it to the work order with the capture time. Photos taken everywhere else are optional.
- 7. Capture the signature and close the job. The customer signs on the device, the service report is generated from the completed record, and completing the booking locks the evidence against further editing by the field role. With no signal, all of it queues and syncs later with its captured times.
How do you make captured evidence hold up in an audit months later?
Capturing evidence is half the job; the other half is being able to show it has not been changed since. Dataverse gives you the controls, but they have to be configured deliberately before the audit rather than after it.
- Enable Dataverse auditing on work orders, work order products, bookings, customer assets and the columns that carry evidence, with a retention period agreed with compliance.
- Remove edit rights on completed work orders and Used product lines from field and dispatch roles; route corrections through a role that has to give a reason.
- Store the generated service report on the work order as a document, so the version the customer signed is preserved alongside the live record.
- Keep captured time and received time separate on records created offline, so late syncs are explainable rather than suspicious.
- Make the history readable for the people answering the question: our free Easy Audit PCF control shows how a value got to where it is, step by step, and exports the filtered history to CSV.
What can you reconstruct when the evidence was never captured?
Honestly, less than people hope, and presenting a reconstruction as original evidence does more damage than the gap itself. What you can do is assemble what survives from independent sources, state how each fact was established, and keep reconstructed data visibly separate from data captured at the time.
| Fact | Often recoverable from | Usually not recoverable |
|---|---|---|
| That a visit happened and when | Bookings, timesheets, vehicle or travel records, emails and calendar entries | Exact arrival and completion times if statuses were never changed |
| Who attended | Bookings, timesheets, expense claims | Who did the work when a crew attended together |
| Parts consumed | Stock movements, purchase orders, returns of defective parts, van stock counts | Which unit on a multi-asset site received the part |
| Serial numbers fitted | Supplier dispatch records, a physical inspection of the asset now | Serials of parts that have since been replaced again |
| Work performed and findings | Photos in email or messaging, customer correspondence, later inspection | Readings and measurements that were never written down |
| Customer acceptance | Signed paper sheets, acknowledgement emails | Acceptance of work that was never confirmed at the time |
How do you document the evidence gap for an auditor or warranty assessor?
The credible position is transparency: here is what we captured at the time, here is what we reconstructed and from what, and here is what we cannot establish. In Dataverse we model this explicitly rather than editing old records to look complete.
- Create reconstructed records, never backdate them to look original: a reconstruction flag, the source of each fact, who reconstructed it and when, and a confidence level.
- Keep the original gap visible in views and reports, so reconstructed and captured records can be counted separately.
- Attach the source evidence (the stock movement, the email, the scanned sheet) to the reconstructed record.
- Write a short gap statement per affected period: the scope, the method, what could not be established and the change that prevents it recurring.
- Where a physical inspection establishes the current state of an asset, record it as a new, dated inspection rather than as the historical visit.
Should proof of work run on Dynamics 365 Field Service or a custom application?
We recommend the right solution - whether that's Microsoft Dynamics 365, Power Platform, or a custom-built CRM. Some businesses need the Microsoft ecosystem. Others need full control without licensing. We deliver both. If your work orders, assets and scheduling already run on Dynamics 365, Field Service is the natural home for the evidence model above. In our case study on Field Service digitization for a European manufacturer, more than 3,000 pieces of equipment were modelled in Dataverse with serial numbers, warranty information and maintenance schedules, and technicians capture photos, measurements and customer signatures in a mobile canvas app. A custom application on React, Node.js, PostgreSQL or .NET is the better answer when field staff or subcontractors are too numerous for per-user licensing, when proof of work has to be delivered inside your own customer portal, or when you need full control of how evidence is stored and retained.
Should proof of work run on Dynamics 365, Power Platform or a custom-built application?
Can afford licensing and want the Microsoft ecosystem
Dynamics 365
Microsoft 365, Teams and Outlook integration, a mature partner ecosystem, Copilot, and apps for sales, service and field operations that are configured rather than built.
Need full control and zero licensing
Custom CRM
A CRM built on React, Node.js, PostgreSQL or .NET that you own outright: your data model, your hosting, no per-user subscription, and features shaped exactly to your process.
Not sure which fits
We help you decide
A short discovery weighs licensing budget, process complexity, integrations and long-term ownership, then recommends one path. We deliver both, so the recommendation has no reason to lean.
What does closing a proof of work gap with Solzet involve?
Solzet has 8+ years of Dynamics 365 Customer Engagement and Power Platform delivery and works remotely from Yerevan, Armenia. A proof of work engagement usually starts with a review of one real audit or warranty question against the live data: which rows in the tables above could be answered from captured records, and which could not. From that we adjust the data model, the mobile capture flow and the security and audit configuration, add completion checks, and, where records are already missing, build the reconstruction model and help assemble the evidence. If evidence gaps are one symptom of a larger Field Service programme across old versions and mixed systems, see Field Service transformation on a legacy estate.
What do people ask us?
What evidence does a warranty claim or audit need from field service work?
Which asset was serviced, its serial number and the serials of components fitted or removed, which parts were consumed, who attended, when they arrived and finished, what was done and found, and proof such as photos and a customer signature. Evidence captured during the visit carries far more weight than records assembled afterwards.
What is the difference between Estimated and Used work order products in Field Service?
Estimated lines record the parts expected for a job, usually added from the incident type. Used lines record parts actually consumed, and can reference the booking on which they were used. Used lines drive inventory adjustments and invoicing, so Estimated lines never switched to Used are a common reason parts are missing from billing and from warranty evidence. There is no separate standard booking product table.
How do we capture serial numbers in Dynamics 365 Field Service?
Hold the serial on the customer asset, with components as child assets that carry their own serials, adding a serial column where your version does not already provide one. On the device the technician confirms the serial with a tap or scans it from the rating plate, and records the serial of any serialised part fitted and removed so the installed base reflects what is in the machine now.
Does capturing proof of work slow technicians down?
It should not if evidence is a side effect of the work. Booking status changes timestamp the visit, scanning the asset opens the job, Estimated parts are confirmed as Used by scan, and photos are requested only at defined points. The only genuinely new actions are those photos and the customer signature.
Can we reconstruct missing service records for an audit?
Partly. Visits, attendance and parts can often be established from bookings, timesheets, stock movements, purchase orders and correspondence, but exact times, readings never written down and which unit received a part are usually lost. Reconstructed records should be flagged as reconstructed, list their sources and never be backdated to look original.
How do we prove a service record has not been altered?
Enable Dataverse auditing on the work order, work order product, booking and customer asset tables, remove edit rights on completed jobs from field and dispatch roles, and keep the generated service report on the record. A readable audit view such as the free Solzet Easy Audit PCF control shows each change with the user and time.
Does Solzet implement the ERP that holds our stock and invoices?
No. Solzet delivers Dynamics 365 Field Service, Power Platform and custom applications, and integrates them with your existing ERP for stock and invoicing. We do not implement Dynamics 365 Finance, Business Central, Dynamics NAV, Finance and Operations or other ERP systems.
Where should you go next?
Dynamics 365 Field Service implementation partner
The asset model, incident types, mobile design and feasibility sprint behind a Field Service build.
Case study: Field Service for a European manufacturer
Customer assets with serial numbers and a mobile app capturing photos, measurements and signatures.
Solzet Easy Audit
A free PCF control that makes Dataverse audit history readable and exportable.
Offline field apps that do not lose data
Keeping evidence captured with no signal safe until it syncs.
Field Service transformation on a legacy estate
A multi-phase Field Service programme across old versions and mixed systems.
Custom CRM Development
Field service and evidence capture on React, Node.js, PostgreSQL and .NET without Microsoft licensing.
Which solution is right for your business?
Tell us what you need. A senior consultant replies within one business day with a recommendation - Dynamics 365, Power Platform, or a custom-built CRM - not a sales script.