The field team can do its job well. The office can do its job well. The operation can still fail in the gap between them.

That gap is where a photograph loses its context, a status changes in one system but not another, or an exception is settled in a message and never returns to the job record. Each team appears productive. Management is left without a reliable answer to a simple question: what can happen next?

A handoff is a management control, not a transfer button. It should give the next person a recognisable job, usable evidence, an operational status, a named exception owner and an explicit next action. When one part is missing, the workflow has transferred uncertainty rather than work.

Follow one real job

Begin with a recent job, not a process diagram drawn from memory. Follow it from planning to site and back to the office. Choose an ordinary case first, then repeat the exercise with one that went offline, arrived incomplete or needed approval.

Look at the records people actually used:

  • the job, asset or work-order identifier;
  • the instruction and evidence requirements sent to the field;
  • location, access and timing information;
  • photographs, measurements, notes or forms returned;
  • status changes in each system;
  • messages used to explain an exception;
  • the office decision and the record of who made it.

The management question is not whether the field form was submitted. It is whether the next person received enough controlled evidence to make the next decision without reconstructing the work from private messages.

Define the handoff contract

A dependable handoff has six parts. Together they form a contract between the field activity and the office decision. The contract states what must be returned, what makes it acceptable and what happens when it is not.

Job identity

Every item crossing the handoff needs a stable reference that both sides recognise. Names, addresses and map locations can help a person recognise the work, but they are weak joins between systems. They change, repeat and invite interpretation.

Choose the identifier at the grain of the decision. If the office approves a work order, return the work-order reference. If it decides by asset, carry both the job and asset identifiers. A photograph named with an address is not reliably attached evidence when several visits or assets share that address.

Status

A status should describe the operational position, not simply that somebody pressed a button. “Submitted” may mean the field team has finished its part. It does not mean the evidence is complete, accepted or ready for the next stage.

Define each important state in ordinary language. State who may enter it, what evidence is required, which transition follows and whether the change is provisional or authoritative. Useful distinctions often include field complete, received, under review, accepted, correction required, revisit required and cancelled.

Evidence

State what evidence is required and what makes it usable. A photograph may need the job reference, evidence type, capture time and location. A measurement needs its unit, method and applicable asset. A declaration needs the person responsible and the version of the question they answered.

Preserve the original item as well as any office interpretation. A caption, classification or extracted value should point back to the source rather than replace it.

Time

Record when the work happened, when the evidence was captured, when the device sent it and when the office received it. These events are often different.

Time matters when a permit, access window, service level or customer commitment can expire. It also locates delay. A capture-to-transfer gap points to connectivity or device behaviour; a received-to-decision gap points to the office queue or missing authority.

Exception owner

When the normal route fails, name the role that acts next. “Send back to operations” is a destination, not ownership.

The owner needs the original evidence, the failed rule, the current state and the action required. They also need authority: a coordinator may request a clearer photograph while only a manager can accept an exception to a controlled requirement.

Next action

The handoff should make the next action explicit. The office may accept the work, ask for a correction, arrange a revisit, seek approval or stop the route. A status that describes the past without directing the next decision leaves the workflow unfinished.

Name the source for each handoff fact

Create a small source-authority map for the handoff:

  • the planning system may own the job instruction and planned date;
  • the field device may own the original capture and capture time;
  • an asset register may own asset identity and location;
  • the office workflow may own validation and the final operational state;
  • a named manager may own an approved exception.

When two sources disagree, apply an agreed precedence for that fact or preserve the conflict for resolution. The latest timestamp is not automatically the authoritative record.

The Government Data Quality Framework treats completeness, uniqueness, consistency, timeliness, validity and accuracy as distinct dimensions. A handoff can pass one and fail another: the evidence may arrive quickly but be attached to the wrong job, or be valid in form but incomplete for the decision.

The W3C provenance model gives a useful structure for retaining entities, activities and agents. In practical terms, keep the evidence, what happened to it and the people or systems responsible for each change.

Control offline and unavailable-system work

Decide before the device loses connection:

  1. which work and reference data must be available locally;
  2. which actions can be recorded offline and which must wait;
  3. how local records are protected and uniquely identified;
  4. how the user can see what is pending transfer;
  5. what happens when the server record changes before synchronisation;
  6. when an unsynchronised job becomes an exception;
  7. who reconciles work that cannot be transferred automatically.

The field worker should be able to distinguish “saved on this device” from “received by the office”. The office should be able to see that expected evidence is still pending. Both views matter; neither is proof of the other.

An unavailable office system needs the same discipline. If somebody records an approval in a temporary channel, retain the job reference, decision, authority, time and evidence basis. Return that decision to the controlled record when service is restored. Otherwise the workaround becomes an invisible second system.

Keep duplicate and out-of-sequence evidence visible

Mobile retries, repeated submissions and delayed synchronisation can make valid events arrive twice or in the wrong order. A simple “last write wins” rule can replace a correct status with an older one.

Give each submission or event a stable identifier. Retain capture, transfer and receipt time to distinguish a delayed original event from a genuinely new one. Deduplicate by identity and business meaning.

Where sequence matters, define the valid transitions. A job accepted by the office should not quietly return to “field complete” because an older device event arrived later. Hold the event, show the conflict and give a named role the evidence needed to reconcile it.

The SIREN portfolio record shows how source health, identity gaps and uncertainty can remain visible in an investigation. It is private Strataflow portfolio work using controlled evidence, not a published client outcome.

Reconcile the office decision

The handoff is complete only when the office decision returns to the job record.

Reconciliation should answer:

  • which field submission the decision addressed;
  • which evidence was accepted, corrected or rejected;
  • who made the decision and under what authority;
  • what operational state changed;
  • whether the field team needs to act again;
  • whether any temporary or duplicate records remain unresolved.

Use a worked handoff contract

The following is a reusable operating template, not a client record. Complete it for one decision before designing screens or integrations.

Scroll horizontally to view all columns.

Contract elementDefinition to agreeExample form of the answer
Job identityThe authoritative reference and the grain of the workWork-order ID plus asset ID
Current statusThe position the field team is assertingField complete, awaiting validation
Evidence packageRequired items and acceptance rulesTwo typed photographs, measurement with unit, signed declaration
TimeCapture, transfer and receipt eventsCapture time from device; receipt time from office service
Source authorityThe system or role owning each factPlanner owns instruction; office validator owns acceptance
Exception ownerThe role acting when a rule failsField supervisor for evidence correction
Next actionAction, owner and acceptance conditionValidator accepts or returns with one named correction
Decision recordWhat must return to the controlled jobDecision, reviewer, time, evidence version and reason

Test the template with a clean job, an incomplete job, an offline job and a conflicting status. If the answers change by circumstance, state the rule rather than relying on local judgement.

Measure whether the handoff is controlled

Management needs measures that expose where the contract fails, not merely how many forms were submitted.

Scroll horizontally to view all columns.

Management questionUseful measure
Did the office receive a decision-ready handoff?Share accepted without clarification or evidence repair
Where is work waiting?Age by handoff state and named exception owner
Does offline work return safely?Count and age of unsynchronised or unreconciled jobs
Is evidence being corrected repeatedly?Correction cycles by evidence type and failure reason
How long does the decision take?Field-complete to office-decision time, with transfer and queue time separate
Does the record contain the decision?Decisions missing reviewer, authority, evidence version or reason

A low exception count may mean the control is working or that failed cases are hidden. Reconcile a sample of accepted work against the original field evidence.

Common failure patterns

Workarounds can make weak handoffs look normal:

  • a photograph is attached but cannot be tied to the required asset;
  • the field system says complete while the office tracker still says planned;
  • a supervisor approves an exception in chat, but the job record never shows the decision;
  • a form rejects an incomplete submission without naming the required correction;
  • an older offline event overwrites a newer office decision;
  • the office rebuilds the job history from email because the systems do not preserve it.

These situations identify the rule, integration or responsibility that needs repair. The smallest useful change may be a stable identifier, a visible pending state or a controlled route back from the office.

Design the exception route

The normal handoff is only half the workflow. Define what happens when the job reference is missing, required evidence cannot be captured, two systems disagree, the device cannot connect, the evidence is ambiguous or nobody present has authority.

Keep the failed condition with the job. Route it with the original evidence and failed rule. Record who accepted, corrected or rejected the case. When the issue is solved outside the main system, return the decision and its basis to the controlled record.

Automation becomes useful after these rules exist. Until then it moves ambiguity faster and makes the resulting backlog harder to explain.

A field-workflow review checklist

Use one real job to answer:

  1. Can both teams identify the same job without interpretation?
  2. Does each status have one agreed operational meaning?
  3. Is required evidence defined and testable at the next decision?
  4. Are capture, transfer and receipt times visible?
  5. Can the office trace each interpretation back to the source item?
  6. Does a failed handoff identify an owner, authority and required action?
  7. Are offline and unavailable-system cases controlled?
  8. Can duplicate or out-of-sequence events be reconciled without losing history?
  9. Does the office decision return to the job record?
  10. Can management see where work is waiting and why?
  11. Can the team explain the complete history without searching private messages?

The Field Workflow Review follows this route over five days. Field Systems & Geospatial covers the system or integration work when the review shows that a build is justified.