# Field-visit completion worksheet

Companion guide: https://overskill.com/learn/plan-field-visit-completion

Copy this Markdown file into your working document. Replace the fictional
service, roles, conditions and times before using it as an app brief. This is
an editable planning exercise; no live field app, inspection certification,
photo storage, location capture or offline behavior is demonstrated.

## 1. Define one service

| Decision | Fictional Cedar Desk policy | Your policy |
| --- | --- | --- |
| Agreed task | Check one meeting room's setup against plan R1 | |
| Visit identity | V-101 through V-104; unchanged across return attempts | |
| Attempt identity | Visit ID plus A01, A02 and so on | |
| Clock | Whole-second UTC timestamps | |
| Current states | Assigned, Attempted, Blocked, Completed | |
| Completion | Current unblocked attempt; exact four passing observations with nonblank notes; summary; worker attestation and submission time | |
| Photo policy | Optional supporting context; never a substitute for any required item | |
| Blocker | Reason, one owner, one next action and future due time | |
| Follow-up timing | Open work is overdue strictly after its due time | |
| Return | Resolve the blocked follow-up, retain old attempt, start an empty next attempt | |
| Duplicate action | Same action reference and entire contents: no change; different contents under that reference: conflict | |
| After completion | No reassignment or edits in this example; define a separate correction policy if needed | |

## 2. Set the required evidence

Each observation stores its own value and nonblank note under the stated key.
Four arbitrary answers do not qualify. Missing values and N/A do not waive a
requirement. Record No or a below-threshold count honestly; completion remains
unavailable until the unmet condition is addressed or a blocker is recorded.

| Key | Value accepted when recording | Passing completion value | Required note |
| --- | --- | --- | --- |
| access | Yes or No | Yes | How authorized entry was obtained; no keys or access codes |
| layout | Yes or No | Yes | Six chairs at the meeting table match plan R1 |
| cups | Whole number of zero or more | 20 or more | Number of paper cups and where counted |
| departure | Yes or No | Yes | Agreed closing procedure completed |

Add a visit summary, worker identity, attempt start and submission time. The
worker attests: “I recorded these observations for this attempt.” A photo does
not establish property safety, substitute for an answer or confirm a successful
upload. Any actual files and permissions need separate tests.

## 3. Reconstruct the October 12 review

All dates are in 2026 and all times are UTC. Dispatch assigns all four visits
at October 12, 08:00. The review time is October 12, 12:00.

| Visit | Worker | Site | Planned start | Actual attempt | Evidence at review | Expected state |
| --- | --- | --- | --- | --- | --- | --- |
| V-101 | W-01 | Alder / room A | Oct 12, 09:00 | A01, started 09:00 | Four passing observations; summary; worker submitted 09:25 | Completed |
| V-102 | W-01 | Birch / room B | Oct 12, 10:00 | A01, started 10:00 | Blocked 10:05; agreed key unavailable | Blocked |
| V-103 | W-02 | Cedar / room C | Oct 12, 11:00 | A01, started 11:00 | Three passing observations; departure missing; one optional photo label | Attempted |
| V-104 | W-02 | Dogwood / room D | Oct 13, 11:00 | None | Assignment only | Assigned |

V-101's filled record, entered at 09:20:

| Key | Value | Note |
| --- | --- | --- |
| access | Yes | Authorized entry using the agreed key. |
| layout | Yes | Six chairs placed at the meeting table, matching plan R1. |
| cups | 24 | 24 paper cups counted in the supply tray. |
| departure | Yes | Door secured using the agreed closing procedure. |

Summary: Room setup check finished against plan R1. Photos: none. W-01 submits
the attestation for V-101-A01 at 09:25.

V-103's record, entered at 11:20:

| Key | Value | Note |
| --- | --- | --- |
| access | Yes | Authorized entry using the agreed key. |
| layout | Yes | Six chairs match plan R1. |
| cups | 20 | 20 paper cups counted in tray. |
| departure | Missing | Missing |

Summary: Departure check still needs recording. Optional photo label: P-103
(fictional, no file uploaded). W-02's completion request at 11:25 is rejected;
the record remains Attempted with no completion time.

## 4. Work the blocked visit and return

| Field | V-102-A01 record |
| --- | --- |
| Reason | Agreed key was unavailable; no authorized entry. |
| Recorded by / at | W-01 / October 12, 10:05 |
| Follow-up owner | COORD-01 |
| Next action | Arrange authorized key handoff and record the return time. |
| Due | October 12, 14:00 |
| Resolved by / at | COORD-01 / October 12, 14:30 |
| Resolution evidence | Property contact confirmed authorized key handoff for October 13, 09:00; fictional reference K-102 |

At 14:00 it is due and not overdue. At 14:00:01 it is overdue. At 14:30 the
follow-up is closed, 30 minutes late; V-102 remains Blocked until a return
attempt starts. Preserve the original due and resolution time.

W-01 starts V-102-A02 at October 13, 09:00 with empty observations. A delayed
A01 completion submission at 09:01 is rejected without changing A02. W-01
records the following at 09:20:

| Key | Value | Note |
| --- | --- | --- |
| access | Yes | Authorized key handoff K-102 completed. |
| layout | Yes | Six chairs match plan R1. |
| cups | 22 | 22 paper cups counted in tray. |
| departure | Yes | Door secured using agreed closing procedure. |

Summary: Return visit completed all four setup checks. Photos: none. W-01
submits the attestation at 09:25. A01 retains its blocked history; A02 holds the
completion record.

## 5. Answer key and cases to test

| Review time | Visits | Attempts | Assigned | Attempted | Blocked | Completed |
| --- | --- | --- | --- | --- | --- | --- |
| October 12, 12:00 | 4 | 3 | 1 | 1 | 1 | 1 |
| October 12, 14:30 | 4 | 3 | 1 | 1 | 1 | 1 |
| October 13, 10:00 | 4 | 4 | 1 | 1 | 0 | 2 |

For each independent completion case below, start with a fresh current attempt
with all other requirements satisfied. A failed action must leave the existing
record unchanged. These are expected planning answers; fill the last column
only after testing your actual app.

| Case | Change or action | Expected answer | Actual app result |
| --- | --- | --- | --- |
| F01 | Assignment only | Assigned; zero attempts, zero completions | Not tested |
| F02 | Four passing observations, summary and worker attestation | Completed once | Not tested |
| F03 | Missing departure plus an optional photo | Reject completion; Attempted | Not tested |
| F04 | Required note contains only spaces | Reject completion; Attempted | Not tested |
| F05 | Layout answer is No with a detailed photo | Reject completion; Attempted | Not tested |
| F06 | 19 cups, then separately 20 cups | 19 fails completion; 20 passes when all else is satisfied | Not tested |
| F07 | Cup count is negative, fractional or a yes/no value | Reject the record update; preserve earlier data | Not tested |
| F08 | An extra or substituted observation key | Reject unknown key; never use it to replace departure | Not tested |
| F09 | Blocker omits owner, action, reason or future deadline | Reject blocker update; no incomplete Blocked record | Not tested |
| F10 | Follow-up reaches exactly 14:00, then 14:00:01 | Due at equality; overdue strictly after | Not tested |
| F11 | Resolve follow-up with owner and evidence | Follow-up closes; visit remains Blocked | Not tested |
| F12 | Start return after resolution | New empty A02; A01 evidence retained | Not tested |
| F13 | Delayed A01 action after A02 starts | Reject; A02 unchanged | Not tested |
| F14 | Repeat an accepted action reference and identical contents | No change, no extra attempt or completion | Not tested |
| F15 | Same action reference with changed worker, time or contents | Conflict; original record retained | Not tested |
| F16 | Different worker submits, or attestation is missing | Reject completion | Not tested |
| F17 | Reassign or edit a completed visit in this example | Reject; use a separately defined correction process | Not tested |
| F18 | Record says N/A or attempts to reuse old observations | No waiver; current attempt needs its own required evidence | Not tested |

## 6. Copy these blank records

Visit:

| Visit ID | Site | Worker | Planned time and timezone | Current attempt | Current state |
| --- | --- | --- | --- | --- | --- |
| | | | | | |

Attempt:

| Attempt ID | Started by | Started at | Required item | Value | Observation note |
| --- | --- | --- | --- | --- | --- |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |

Summary: ______. Worker attestation: ______. Submitted by/at: ______.
Optional supporting photo references, if separately implemented: ______.

Blocked follow-up:

| Attempt | Reason | One owner | Next action | Due and timezone | Resolved by/at | Resolution evidence |
| --- | --- | --- | --- | --- | --- | --- |
| | | | | | | |

Action history:

| Action reference | Visit | Attempt | Action | Actor | Time | Contents or revision reference | Accepted / rejected / retry |
| --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | | | | |

Keep rejected actions out of the accepted state. In a real app, decide how to
retain a separate diagnostic history without exposing private details. The
worksheet's action references describe a retry rule, not tested delivery,
authorization, simultaneous writes or durable storage.

## 7. Add the missing implementation decisions

- Which account can read or change each visit, direct link and attachment?
- What happens if the worker leaves, refreshes or loses a connection before submission?
- How are corrections, cancellations and reassignment reviewed and retained?
- If photos are added, which file types, sizes, permissions and save/reload checks apply?
- Which phone sizes and keyboard paths have been tested with the longest notes?
- What observed result proves the worker can recover from each failed action?

Use the [app brief guide](https://overskill.com/learn/write-an-ai-app-brief),
[design review](https://overskill.com/learn/make-an-app-easier-to-use),
[private-access checks](https://overskill.com/learn/verify-private-record-access)
and [launch checklist](https://overskill.com/learn/ai-app-launch-checklist) to
turn those decisions into checks. Keep each result marked Not tested until
someone verifies it through the actual app.