Skip to content
MirrorFirm

What does one generated task look like for a utility connection request?

A task pack for a utility connection request carries starting state, a policy rule, a scripted late document and an expected end state, which together define pass or fail.

A utility's customer service team receives a new electricity connection request, CASE-0412, for a rural property. The environment state for this task holds the case, a party record for the landholder, a site record SITE-8841 with a distribution zone, three documents, a policy rule for connections above a stated load, and two open tasks.

The agent's goal is to bring the case to a routable state. It lists documents and finds a site plan and an electrician's certificate but no landholder consent, which the rule requires because the party record shows a lessee rather than an owner. It calls request_information for the consent, sets the status to PENDING_INFO, and attaches a note quoting the rule identifier.

When the consent arrives as a scripted event in the task, the agent reopens the case, checks the document type, and routes it to the network engineering queue rather than the standard connections queue, because the load exceeds the rule threshold.

**Consequence:** the final snapshot shows status ROUTED, queue network-engineering, one information request closed, and no change to the site record. A grader reads those four facts directly.

**What it teaches:** a task pack carries starting state, scripted events, a policy that drives the decision, and an expected end state that names the fields that must stay untouched.

A task is more than a prompt: starting state, scripted events, a governing rule and an expected end state make grading mechanical.

Audience:
Environment builders
Content type:
Scenario
Buyer question:
What can be generated

Content release: r-a220eb6e099b243b

More on Application intake