System Online|Autonomous Mode
CH-04 // Family Channel

You cannot see the group chat. That is the whole problem.

The most common place a child is cyberbullied is an online group chat, where no parent, teacher or moderator is present. Guardii is built to detect the pattern inside that layer and tell you what is happening, while leaving your child the privacy that keeps them using their phone honestly.

Startmatebacked
Official Meta Business Partner
// Section 01 · What the data says

You are looking. It is happening where you cannot see.

Every figure below is from the eSafety Commissioner, the Australian Government agency for online safety.

39%

of cyberbullied children were last targeted in an online group chat, the most common setting of all. Only 8% were last targeted on someone else’s social media page.

eSafety Commissioner, 2024–25 (n=696 cyberbullied children)
51%

of parents whose child had been treated hurtfully online were aware it had happened, even though 69% of those children said they told them.

eSafety Commissioner, Mind the Gap (2021 fieldwork)
13%

of Australian children aged 10 to 17 have been told online to hurt or kill themselves, or that they should die.

eSafety Commissioner, 2024–25 (n=3,454)
→ Read the full research
// Section 02 · What Guardii looks for

Patterns, not keywords

Sustained targeting

The same person coming back again and again. A third of cyberbullied children said their most recent experience was repeat behaviour by the same person.

Exclusion and pile-ons

Group chats that turn on one child, or a child quietly removed and talked about. Being deliberately left out of an online activity was the second most common form of cyberbullying children reported, and it is among the hardest to see.

Distress signals

Language indicating a child may be in crisis, including being told to hurt themselves, which 13% of cyberbullied children reported. Detection supports a parent noticing sooner. It is not a clinical or crisis service, and it will not catch everything.

Grooming and sextortion patterns

Escalation sequences, isolation attempts and coercion, detected by structure rather than by individual words.

// Section 03 · How it works

Three steps, and your child keeps their privacy

STEP 01

Connect your child’s accounts

You and your child connect the accounts you agree to cover. Guardii is not spyware installed behind their back, and it works best when the child knows it is there.

STEP 02

Guardii looks for harm, not for content

Detection looks at the shape of a conversation over time, including repetition, escalation and coordinated targeting, rather than scanning for banned words. Platform coverage depends on what each service permits, and we will tell you exactly what is covered before you sign up.

STEP 03

You get told what matters

When a pattern crosses the line, you get an alert describing what is happening and how serious it is. You do not get a feed of your child’s conversations.

// Australia · the two school day rule

Schools will have two school days to start acting. The clock starts when they find out.

By Term 1 2027 all Australian schools, government, Catholic and independent, will publish or link to an anti-bullying policy aligned to the National Framework or an equivalent framework in their state. It expects schools to make reasonable effort to initiate a response within two school days of becoming aware of the behaviour. A school cannot act on a group chat it has never seen, so how you report matters.

// Section 04 · Questions parents ask

Questions parents ask

Does Guardii let me read my child’s messages?

No. Guardii reports patterns rather than transcripts. You are told that sustained targeting or a grooming pattern has been detected, how serious it is, and on which platform, so you can act. You do not get a running feed of your child’s private conversations, because a tool that removes a teenager’s privacy is one they will simply route around.

Why can I not just check my child’s phone?

Because most of it is gone before you look. Messages disappear, group chats scroll, and children delete what they are ashamed of. In eSafety’s 2024–25 survey the most common setting for a child’s most recent cyberbullying was an online group chat, named by 39%, against 8% who were last targeted on someone else’s social media page. In eSafety’s earlier Mind the Gap study, only 51% of parents whose child had been treated hurtfully online were aware it had happened.

My child will not talk to me about it. Is that normal?

It is very common, and it becomes more common with age. Among children who had been treated hurtfully online, eSafety found 64% of 14 to 17 year olds told a parent, compared with 75% of 8 to 10 year olds. Disclosure falls as children get older, which is part of why noticing it yourself gets harder over time.

What platforms does it cover?

Coverage depends on what each platform permits a safety tool to access, and it changes as those rules change, so we confirm the current list with you before you sign up rather than promising it here. For context, the services Australian children named most often for their most recent cyberbullying were Snapchat (21%), online video games (20%), Messenger Kids (13%), Instagram (8%), Facebook (7%) and Discord (6%). Note also that since 10 December 2025 several of these platforms are age-restricted in Australia for under-16s.

What should I do if my child is being bullied at school?

Report it to the school in writing. From Term 1 2027 Australian schools will publish or link to anti-bullying policies aligned to the National Framework, which expects schools to make reasonable effort to initiate a response within two school days of becoming aware of the behaviour. A written report starts that clock and creates a record. Guardii publishes a free letter template you can use.

// Section 05 · Where we draw the line

Safety and surveillance are not the same thing

One watches people. The other watches for harm. Most tools sold into this category blur that line, and every time one does, it makes the people it was meant to protect less safe, because they route around it. Guardii is built so the distinction holds under pressure.

Patterns, not transcripts

Detection reports the shape of a conversation — sustained targeting, escalation, coercion — not its contents. You are told what is happening and how serious it is. You are not handed a feed of what was said.

Nobody reads it, including us

Content is committed in a form we cannot browse or mine. We hold the integrity of the record rather than access to it, which is what lets the same system protect a person and preserve evidence without reading either.

The person protected knows it is there

Guardii is not installed behind someone’s back. A child, a student or a worker is told what is covered. Protection people discover rather than consent to is the thing that destroys trust, and trust is what makes them keep talking.

The subject is the client

Alerts serve the person at risk and the adult responsible for them. Where an organisation pays, it receives what it needs to run the programme, never a window into somebody’s private life.

Least data, shortest path

We take the minimum required to detect the harm, and we do not repurpose it. No profiling for advertising, no resale, no secondary use dressed up as product improvement.

This is a design constraint rather than a policy preference. A system that cannot read what it protects cannot be quietly repurposed into one that does, by us or by anyone who acquires us.

If you need help now

If your child is in immediate danger, call 000. Kids Helpline is free and available to young people 24 hours a day on 1800 55 1800, and Lifeline is on 13 11 14. Seriously harmful online content targeting an Australian child can be reported to the eSafety Commissioner once it has first been reported to the platform it appeared on. Guardii is a detection tool and is not a crisis service.

Get early access to Guardii Family

Guardii Family is rolling out to Australian households. Leave your email and we will tell you when it opens, along with the practical guidance we publish for parents. No spam, and you can leave at any time.