Beta testing for AI/ML

Test your new tech to make sure it works for users.

Before you release your products and services that use data, AI or ML, it's important to work with users to make sure its trustworthy.

Outcome

  • manage human-centred risks that come with using AI/ML in your products or services

  • help your customers understand and trust the AI/ML technology that you're using

  • guide your organisation to use technology in an ethical way.

Service description

Many new products and services use new technology and data, like AI/ML, to add value for customers.

But people often don't fully know how the AI/ML affects their experience when they use your products and services.

It's even harder for customers to tell you what they think about these technical parts of the product or service compared to other user experience features.

You would usually then do rounds of UAT where stakeholders and customers assess and give feedback on the product or service before launch. These UATs usually don't focus on the AI/ML components specifically – that’s where we come in with Beta Testing.

Our approach

Data Wisely applies the same user-centric mindset of UATs to the AI/ML components of a product or service. Our service to you:

  • Test and report on how acceptable it is to users.

  • Find any critical risks that come with using AI/ML.

  • Recommend ways to fix issues and improve the user experience.

  • Give you advice on how to internally manage the product or service.

Engagement

Depending on how complex your product or service is and the plan we’re using, Beta testing usually takes between 3 and 6 weeks.

We tailor our approach to your needs and are sensitive to what they are.

This is what a typical partnership might look like when you engage us for AI/ML Beta testing.

Initial briefing meeting (one hour):

  • Meet with the core project team or product owner to get to know each other and go over the project goals and plan.

One week of collaboration:

  • Try the product or service out in its test state.

  • Collaboratively identify the key research questions and acceptance criteria to test.

  • Agree on any ethical tech principles or policies that the organisation needs to uphold.

  • Identify the participants for testing and recruit them.

Two weeks of testing:

  • Design and conduct the tests.

Two weeks of reporting and feedback:

  • Synthesise, analyse, and report on the test findings.

  • Present the findings and initial recommendations to your key stakeholders.

  • Finalise and prioritise the recommendations into an action plan through a stakeholder workshop with decision-makers.

Cost

The cost of UAT usually falls between $10,000 and $50,000 (ex GST).

This varies depending on the product or service’s scale and complexity.

Let’s chat about the best approach that fits in your budget and gives you and your team the most value.