Fictional technician reviews a report beside an air conditioning unit and filters
AI-generated image; illustrative scene.

The visit is over: which filter was cleaned and what is still open?

Prepare HVAC reports by equipment, separating completed and outstanding work with supervised AI.

Adaris Tech Team6 min

The report must separate completed and outstanding work

On a visit with several units, a note such as “filters checked; one pending” does not say which filter was cleaned or which remains unchecked. Before handing over the report, the technician needs to associate each action with the right unit and clarify what remains. A custom solution could prepare a draft with the original notes in view, but must never turn ambiguity into completed work.

An Obsiva template proposes identifying the unit by location, make, model and code or serial number, and recording the final status and next date. RefriFast describes an equipment history with dated completed work and parts fitted; these are organisation examples, not a universal rule or confirmation that maintenance occurred.

Two units, one visit: a fictional case

An imaginary office has two units identified as A and B. The work order asks for both to be checked. The technician notes “filter A cleaned” and “filter B inspection pending because access was unavailable”. The draft shows the completed action for A and the outstanding inspection for B separately, with the note and visit date. It does not infer that B’s filter is clean or assign a return visit without agreement.

If another note says only “filter pending”, the system leaves it unassigned and asks for clarification. The technician checks their notes, corrects the draft and decides what to communicate to the customer. Professional assessment determines the status and necessary work; the report is neither a diagnosis nor permission to operate equipment.

From notes to a report reviewed by a person

Adaris could assess an application bringing together the work order, authorised equipment inventory and visit notes or photos. AI could suggest fields, retain the source excerpt and flag contradictions or missing details. OpenAI documents models that can read visible text in images and recommends enlarging small text: a suggested reading can be wrong and must be checked against the original.

A person checks the unit, location, date, actions, outstanding work and any measurements before approving the report. Preserving context can also help with a roof photo report and open tasks. If a part is needed, ordering, receipt and the appointment are separate states, as discussed in the guide to a return visit awaiting a spare part.

When to integrate and when a record is enough

If the team already works well with a shared record, organising its fields may be enough. When notes and inventory live in different systems, Adaris could assess an integration with limited permissions. A constrained agent could consult authorised records and prepare questions about gaps; OpenAI documents connecting models to application data and functions. This is a proposal for assessment, not an existing Adaris integration.

In the pilot, access would be read-only and the draft subject to review: changing the history, sending the report or scheduling a visit would require separate permissions and decisions. Agree authorisation and processing terms before using real personal data, images or documents.

A pilot the service team can evaluate

Prepare fictional or authorised cases: two units, a completed task, an outstanding task and an unidentified note. A person first establishes the correct reference. Then compare the draft and count tasks attributed to the wrong unit, outstanding tasks presented as complete and uncertainties flagged; also record review time. These are measures for deciding whether the process helps, not promised results.

If the responsible person finds the result useful and can correct it, the next step is to define sources, permissions, validation and connections to actual tools. The technician keeps the decision over the work and the final report. Commercial enquiries are handled in Spanish.

Related services

Sources

Do your reports mix completed and outstanding work?

At Adaris, we specialise in integrating AI into equipment-level HVAC reports, distinguishing cleaned filters, completed work and outstanding tasks for staff review. Describe your process without private data and we can assess a custom solution. Commercial enquiries in Spanish.

Discuss my process

Back to all insights