CONTEXT(CLONE)
Sign in

Check site photos against the construction programme

For: Construction project manager or owner's representative tracking progress across sites

Pattern: Map, verify, reduceNeeds physical AI: plannedDesigned for 30 to 800 agents
Get early accessSign in to use

The pain today

Progress is reported as a percentage in a spreadsheet. Thousands of dated site photos exist but nobody compares them with the programme, so slippage and out-of-sequence work show up at the next valuation.

The ask

“I attached this month's dated site photos by zone, the construction programme and the drawings list. For each zone, tell me what work the photos show as started, complete or absent against the programme, with the photos behind each finding.”

Plain words, as you would say it to a colleague. Edit it to fit your case before you send it.

What you attach or connect

  • Dated site photos by zone
  • Construction programme
  • Drawings list and zone plan
  • Previous progress report

The unit of work

One worker task per one zone's photo set for one period.

Why a swarm fits

Each zone and period is judged from its own photos against a few programme activities. Zones are independent, and only the reconciler needs the whole programme.

Not for

Measuring quantities or certifying payment. Photos show that work is present, not its quality or measured extent.

The decision tree

5 typed decisions, each with an action for every answer

At fixed moments in a run, the engine puts one narrow question to a decision model. The decision model never writes text: it answers yes or no with a probability, picks from listed options, or gives a score, about a small slice of the material. The engine then does exactly what this tree says, which is what makes the run auditable. The thresholds are the template's design values, not measured results.

  1. Planner, while planning

    Scope checkA choice among options

    Before work starts on a unit

    Does the description of this zone's photo set say the photos are dated within the period and show the named zone?

    Sees only: Planned design: one zone and period, with the perception model's written description of the photo set

    Why: Planned: zones without usable photos are reported as unassessed rather than as showing no work.

    • Dated and in zonethenAccept
    • Undated or wrong zonethenSkip this unit
    • Too few photos to judgethenMark unresolved

  2. Before workers, before a task runs

    Small worker or strong modelYes or no, with a probability

    Before a task runs

    Is this programme activity one whose start or completion would be visible in a photograph, rather than concealed or internal work?

    Sees only: One programme activity's description

    Why: Planned: concealed work is marked not observable instead of spending perception effort on it.

    • Yes: 0.60 or higherthenAccept
    • Unsure: 0.30 up to 0.60thenEscalate to a strong model
    • No: below 0.30thenSkip this unit

  3. After workers, the judge checks

    Evidence checkYes or no, with a probability

    After a worker answers

    Does the written description of what the cited photos show support the status given to this programme activity: started, complete or not visible?

    Sees only: Planned design: one activity, its stated status and the description of the cited photos

    Why: Planned: a status must follow from what was described, not from what the programme expects.

    • Yes: 0.85 or higherthenAccept
    • Unsure: 0.50 up to 0.85thenMark unresolved
    • No: below 0.50thenReject and retry

  4. Reconciler, while merging

    Conflict checkYes or no, with a probability

    While reconciling

    Does the observed status of this activity disagree with the status in the contractor's progress report for the same period?

    Sees only: One activity's verified observed status and its line in the progress report

    Why: Planned: the disagreement is shown to the project manager, never silently corrected.

    • Yes: 0.70 or higherthenMark unresolved
    • Unsure: 0.30 up to 0.70thenMark unresolved
    • No: below 0.30thenAccept

  5. Accountable person, before anything is settled

    Person decidesYes or no, with a probability

    Before anything is reported as settled

    Does this finding differ from reported progress or show work out of programme sequence, so that it could bear on a valuation or notice?

    Sees only: One zone's findings with photo references

    Why: Planned: the project manager and quantity surveyor confirm on site before money or notices move.

    Accountable: The project manager and quantity surveyor confirm progress on site before any valuation or notice.

    • Yes: 0.30 or higherthenAsk a person
    • Unsure: 0.10 up to 0.30thenAsk a person
    • No: below 0.10thenAccept

The fleet: who does what

Model tiers by role, not brands: you choose the models. Strong reasoning models plan and reconcile, small fast models do the wide work, and the judge is a decision model from a different family, so it does not share the workers' blind spots.

  1. Planner

    A strong reasoning model turns the programme into visible activities per zone and period.

    Decisions here:1. Scope check

  2. Workers

    A perception model for video and images, planned, one task per zone, describing visible work with photo references.

    Designed for 30 to 800 agents, one worker task per one zone's photo set for one period. Each worker receives only its own unit.

    Decisions here:2. Small worker or strong model

  3. Judge, from a different model family

    A decision model from a different family checks that each described observation supports the stated activity status.

    Decisions here:3. Evidence check

  4. Reconciler

    A strong reasoning model compares observed status with the programme and the reported percentage.

    Decisions here:4. Conflict check

  5. Accountable person

    The project manager and quantity surveyor confirm progress on site before any valuation or notice.

    Decisions here:5. Person decides

Checked before anything is accepted

  • Every status cites the photos it rests on, with date and zone
  • Activities that cannot be seen in photos are marked not observable, not absent
  • Zones with too few or undated photos are reported as unassessed
  • Disagreement with the reported progress is shown, never silently corrected

What comes back

  • Per zone: activities seen as started, complete or not visible
  • Differences from the reported progress, with photos
  • Work seen out of programme sequence
  • Zones without usable photo coverage

What to measure

  • Share of findings confirmed on a site walk
  • Slippage identified before the valuation versus at it
  • Project team hours per progress report
  • Cost per zone assessed

Names of measures only. No result is claimed for this template.

Templates open in the workspace chat with the ask filled in. Nothing runs until you send it.

Get early accessSign in to use

More in Physical AI and field operations

Roll up drone inspection footage across an asset network

For: Asset integrity manager for overhead lines, pipelines or telecom towers

Drone flights return footage for every span and tower.

Pattern: Hierarchical decompositionNeeds physical AI: planned6 decisionsDesigned for 50 to 3,000 agents

Review plant walkdown video from four specialist angles

For: Site HSE manager or operations manager at a process or manufacturing plant

Walkdowns are filmed but reviewed by one person with one checklist.

Pattern: Specialist panelNeeds physical AI: planned5 decisionsDesigned for 8 to 300 agents