Turn explicit field requests into a commercial HVAC sales referral

Capture the request in the customer’s terms

A technician may hear a facility contact ask for a price on additional work during an existing visit. That request can get buried in a service note. This workflow prepares a referral for office review; it does not discover replacement needs or recommend technical work from equipment age. Begin with an authorized note that identifies the location, visit and exact request.

Separate what the customer asked, what the technician observed and what the technician recommended. Ask AI to retain those as different fields. If no customer request appears, the result should say that explicitly instead of generating a sales lead. Any technical concern still goes through your qualified service review process, independent of whether a commercial opportunity is opened.

Give the office enough context to qualify it

Use referral ID, source job/location, request date, supplied business contact, requested work, unanswered questions and assigned reviewer. Include a link to the approved note and state whether photos are relevant and permitted to share. Do not copy access codes, unrelated tenant information or payment details into a sales queue.

In a fictional maintenance visit, a facility manager asks for a discussion about adding another building to the service program. The referral can repeat that request and identify the building if supplied. It cannot conclude that the existing agreement covers the site, that the person can sign, or that a particular visit frequency is appropriate. Those are qualification questions for the account owner.

Confirm ownership before creating commitments

Have an office reviewer check whether an existing opportunity already covers the request. Link rather than duplicate when appropriate, and retain the original field evidence. The sales owner chooses the next qualification step and any approved customer contact; AI should not automatically create a quote, book a site visit or promise a response window.

ServiceTitan’s agreement overview describes linked pull-through jobs, opportunities and estimates. That provides record context, not proof that your account has an automatic field-to-sales AI workflow. Test this proposed queue using fictional notes with explicit requests, ambiguous remarks and no request at all. Reject outputs that turn every service observation into a lead. Track whether handoffs preserve the customer’s actual question before considering a live pilot.

Review checklist

• Require an explicit request or mark it absent.
• Separate request, observation and recommendation.
• Retain the source job and location IDs.
• Check for an existing related opportunity.
• Have the account owner approve the next action.

Source context and limits

ServiceTitan documents pull-through records associated with agreements. The request-extraction queue is our proposed editorial workflow; it is not an assertion of an automatic referral feature or evidence that a customer approved additional work.