How We Work

When you need us,
a real person is in the loop.

We use software to handle the routine work so our people can focus on the interactions that actually need them. This page explains what that looks like in practice — and what we won't do with it.

Three things that make our service different

Here's what we actually do — not what every ISP says about AI, but what we've built and how it works.

Human-in-the-Loop support — AI drafts, humans approve

Inbound email and chat messages go through our Human-in-the-Loop (HITL) model. Phone calls are answered directly by a person. For email and chat: an AI reads your message and drafts a response. A real person at HIT Pacific reads the draft, looks up your account, approves or edits the response, and sends it. The AI helps with the first draft; the human handles the judgment, the account context, and the send.

A defined set of routine, low-stakes responses — outage status updates, service availability queries — are handled automatically without a separate human approval step. These are categories where the response is factual and mechanical: the answer comes from a live data source, not from judgment. Complex, sensitive, billing, or ambiguous situations always require human approval before anything leaves our system. That line is a deliberate policy decision, not a cost-cutting exercise.

We log every AI interaction — what was drafted, what was changed, what was sent. That log is what makes us accountable: if a response was wrong, we can see exactly what the AI produced, what a human approved, and where the process broke down.

Proactive fault management — you shouldn't have to call us

When there's an outage on our wholesale network or the NBN, our systems detect it before most customers notice. Affected customers get an automatic notification. Status updates go out on a defined cadence without anyone needing to write them. The outage log is updated in real time.

For faults that persist beyond same-day resolution, staff takes personal ownership — escalating directly with our wholesale suppliers, keeping customers updated, and working the fault until it's closed. The AI handles the communication cadence; the human handles the relationship.

Your account data and the AI work separately — in our current model

In our current operating model, AI agents at HIT Pacific work from what you tell them in the conversation. They do not pull your personal data from our systems, and they do not access your account records.

Account operations — billing, provisioning, plan changes, cancellations — are handled by purpose-built code, not by AI browsing your account. Code that touches important data is written specifically for that purpose, reviewed, and auditable. That is a deliberate design boundary, not a limitation.

When you contact us, an AI reads your inbound message to understand what you need. It works from what you've written — not from your account records. It can help you understand your options, talk through a problem, or direct you to the right place in our self-service portal — but it does not see your service details, billing history, or personal information unless you share them yourself in that message.

What we won't pretend

AI makes mistakes. Ours will too. The HITL model exists specifically because we know this — for anything that requires judgment, a human reads the AI draft before it leaves our system. For the narrow category of automatic responses, we limit them to: factual replies based on live data (outage status, service availability checks), and brief welcoming messages to prospective customers directing them to our website. These are categories where the response is either mechanical or carries no judgment risk. More complex, sensitive, billing, or ambiguous situations always require human approval before anything leaves our system. We don't claim the AI is infallible. We claim the process around it is designed to minimise the impact of errors before they reach you.

We will never use AI to make it harder to reach a human. If you want to talk to a person at HIT Pacific, you can — by phone on 07 3569 1900 (answered directly by a person), or by email at hello@hitpacific.com.au (goes through our HITL model; include your complaint reference number to flag urgent cases). The AI is there to make the interaction faster and better, not to act as a barrier.

Our AI policy is published in our Standard Form of Agreement (Section 20). If you have a dispute about how AI was used in your support interaction, the complaints process at /complaints/ handles it.