How it works
The experience
01
Something happened. They find us.
They land on isthisokay.org. First thing they see: “You're not sure if it counts. You're in the right place.” No account. No form. No demand before they're ready.
02
They tell us what happened. In their words.
No checkboxes. No legal categories. Plain language. Their language.
They don't translate for us. We translate for them.03
We extract what the law needs to know.
When. Who. What. The power relationship. The system identifies the legally relevant elements, so they don't have to organise their experience into a framework they don't yet have.
04
We show specific legal violations based on local law.
Two independent AI systems check their situation against Australian law and hundreds of real case precedents. Both must agree before anything surfaces. When they agree, the person sees the specific legal provisions that apply to their situation and the precedents that match it. Not general information. Cases that look like theirs. When the systems disagree, they are told there is uncertainty. Not given a confident wrong answer.
05
Clear pathways for action.
Internal escalation. AHRC complaint. Fair Work application. Every option that applies to their situation, presented clearly. None pushed. They decide if, when, and how.
06
They leave with a reporting-ready document.
Whether they act or not, they leave with everything. A fully formatted incident log ready to submit. The specific legal provisions that apply. The precedents relevant to their situation. A clear understanding of every option available to them. No pressure to use any of it. Just everything they need, ready, if they decide to.
Two layers. Not one.
The technology handles what technology is good at.
The community handles what only people can do.
Tech infrastructure
Legal literacy. Documentation. Verified pathways. The information layer that currently only exists for people who can afford a lawyer.
Community scaffolding
Peer workshops. Real connection. The isolation that the system creates is just as damaging as the harassment itself. We are not using AI to simulate human support. We are creating the conditions for it.
Under the hood
Anonymous
No account. No data stored. Session ends, everything goes. Built into the architecture, not a setting.
Dual-validated
Two systems. One answer required. Uncertainty disclosed. Wrong legal information is worse than none.
Trauma-informed
No time pressure. No self-classification. Every interaction designed around Doubt and Denial.
Information, not advice
What the law says. Never what to do. Enforced in the architecture, not just the terms.
Modular legal layers
Phase 1 is Australian law. Adding UK, NZ, EU is an extension. Not a rebuild.
Outcome tracking
Phase 2. Aggregate, anonymised. Is she still in her field 12 months later? Nobody else is measuring this.
2
systems must agree
When both agree, the answer surfaces. When they disagree, the person is told — not given a confident wrong answer. We designed around the failure case, not the success case.
What comes next
Pilot: In Beta
External accountability mechanism
Organisations can no longer claim they didn't know. The moment a complaint is received, the clock starts. It runs publicly.
International expansion
Individual accountability
Digital spaces
If this is your fight too, there's a specific place where your work changes the outcome.