Services

Usability testing and AI assistant reviews.

We help Australian product teams see where customers get stuck with an AI assistant, an AI feature or a digital service, and what to change. Based in Canberra, working with teams across Australia.

AI assistant usability review

We take your assistant or AI feature through realistic customer tasks and show where it helps, where it stalls and where people can’t reach a person. Each finding points to the exchange it came from, with priorities and suggested changes.

Usability testing with your users

Moderated sessions with representative people, recruited for your audience or drawn from your customers. We watch what they understand, expect and do, then report what to change and why.

Ongoing assistant checks

AI assistants change without anyone redesigning them. We repeat the same tasks on a regular schedule, with real people from time to time, so you see when an answer gets worse before customers do.

How we review an assistant

Five things we check in every assistant.

We look at usability the way ISO 9241-11 describes it: can people get the task done, with reasonable effort, and come away satisfied? Automated runs set a baseline. Sessions with real people show the effort, understanding and trust behind the result.

This is experience testing, not functional QA. An assistant can work exactly as specified and still leave a customer stuck.

Discover: Can people tell what the assistant can do, and when to use it?
Ask: Does it understand plain language and the context a customer has already given?
Answer: Is the answer right, clear and quick enough to act on?
Trust: Do people trust it the right amount? We compare how confident they feel with whether the answer was correct.
Escape: Can people reach a person, make a complaint or ask for hardship help when they need to?

A few useful answers

Before you get in touch.

Do you work outside Canberra?

Yes. Most reviews and research sessions run remotely, so we work with teams anywhere in Australia. In-person sessions can be arranged in Canberra or elsewhere.

What kinds of AI experiences can you review?

Customer-facing chat assistants, AI search and help features, and copilots inside a product. We can review a live service or a prototype; questions about real output quality need access to the working AI.

How is this different from QA or bot testing?

Functional testing checks that the assistant works as specified. We check whether a customer can get their task done, with reasonable effort, and trust the result. The two work together, and any bugs we notice along the way go to your team.

How much does it cost?

It depends on the scope. The first conversation is free, and any work that follows is scoped and priced in writing before it begins.

Start with a conversation

Let’s talk about where your users get stuck.

Tell us where people struggle and what you need to decide. We’ll suggest what would help most in a free, 30-minute call.

Book a call