The expensive part of AI is the handoff.

Your team uses more than one model, agent and tool. Work gets lost between them, agents end up with more access than they need, and nobody can quite say what happened last Tuesday. We're building the layer that fixes that. While we finish it, we'll look at your setup and tell you straight.

We haven't launched our product yet. We'd rather say that up front than pretend otherwise.

Agent Stack Audit

For small technical teams (2 to 15 people) doing real work with several AI models.

The audit is meant for teams of 2 to 15. A larger team can still ask.

What you get

  1. A map of your AI stack. Which models, agents and tools touch which work, and where your data goes, by provider and region where the provider says.
  2. Your handoff points. Where people re-explain context to the next model, agent or person, and what that costs you in time and mistakes.
  3. An access review. What each agent or integration can reach, and where that's broader than the job needs.
  4. An evidence check. On one or two of your real tasks, could you show who asked, what ran, which model did the thinking and what changed?
  5. A short written report. Findings ranked by impact and effort, and a call to walk through it.
  6. A straight answer on fit. If Primitive isn't right for you, the report says so.

How it works

  1. Tell us about your team. Fill in the form below. It takes about two minutes.
  2. Intro call. A short call to agree what's in scope. We'll tell you how long it is before you commit.
  3. Walk-through. You show us how work moves through your tools. We work from screenshots, config exports and a screen share. We don't ask for passwords or keys.
  4. Report and walk-through. You get the written report and a call to go through it. We'll agree the timing before any work starts.

Questions

Who does the audit?

The founders of Primitive, a Perth company. We'll tell you who is on the call before it happens.

How much does it cost?

We haven't published a price. We'll confirm it before you commit to anything.

Do you need access to our systems?

No. We work from what you show us. We won't ask for passwords, keys or admin access.

Is this a sales pitch for your product?

Partly, we'd be lying if we said otherwise. But the report is useful on its own, and if Primitive isn't a fit we'll tell you. Our product isn't launched yet anyway.

Will this make us compliant with the Privacy Act or anything else?

No. We don't certify anyone. We'll help you see what your AI does and where your data goes, which is a good start if you need to explain it to someone.

Where does our information go during the audit?

Only the notes and files you choose to share, kept for the audit. How long we keep them isn't settled, and we won't name a country for storage until that has been checked.

What tools do you cover?

Whatever your team actually uses: chat assistants, coding agents, automations and the integrations between them. If we don't know a tool well, we'll say so.

We're bigger than 15 people. Can we still book?

Yes, get in touch and we'll tell you honestly if the audit suits a team your size.

Why should we trust a pre-launch company?

Fair question. We publish what didn't work in our research, and we'll be just as plain about your setup. You can judge the report on its own.

Which AI tools does your team use? optional

We use your details to reply about the audit. A privacy policy isn't published yet.

Join early access

We're building the operating layer between AI and the work it does, so context, permissions and a record of what happened stay intact as work moves between people, models and tools. It isn't ready yet. Leave your details and we'll tell you when it is.

What to expect

  • A note from us when there's something you can try. We haven't set a date.
  • Honest updates on what's working and what isn't.
  • No sales calls unless you ask for one.

We use your details to tell you when there is something to try. Nothing is turned on by this form.

Perth, Western Australia.