At this week’s QA Financial Forum London 2024, Lee Kivell, Architecture and Engineering Director at Vitality, presented a Case Study in AI for Test Automation, in partnership with Qualitest.
During his session, panellists heard how Vitality, a global health and wellness company that encourages positive lifestyle choices among its members, has leveraged AI technologies for automated testing.
Following the conference, QA Financial checked in with Lee.
QA Financial: AI is very much the buzzword at the moment. How has Vitality leveraged AI technologies for automated testing?
Lee Kivell: Vitality had a large amount of data available in relation to the manual testing being performed. So, when we embarked on a strategy to automate as much testing as possible, the use of AI models allowed us to prioritise the areas of most value first. For example, highlighting the tests that would allow us to find defects at the earliest opportunity, or identifying repetition in test packs that could be eliminated.
“The increase of test automation was an important element of a wider test strategy programme, and the use of AI was a critical accelerator.”
– Lee Kivell
QA Financial: How has AI re-defined business-driven testing for you?
Lee Kivell: I wouldn’t say that AI has completely redefined business driven testing, but we have been able to use AI/ML modelling to focus testing in the most effective areas, and supporting the wider “shift left” approach. This has enabled us to ensure that business acceptance is focussed on testing the true business impact of a change, rather than another level of functional defect identification. The use of LLM models is also enabling us to generate test cases more quickly.
QA Financial: So it will push the industry forward, boosting QA capabilities?
Lee Kivell: I think it is important to acknowledge that although AI has become a big buzz word over the last year or so with the success of ChatGPT and other large language models, more traditional AI and Machine Learning have been around for a very long time now.
Both approaches have an important place in testing and the wider software development lifecycle – although the general AI term can cause confusion.
Therefore, I tend to talk internally about “AI for decision making” (AI/ML) or “AI for content generation” (AI/LLM) to help distinguish, I’m sure there are better descriptions in use elsewhere.
“A lot of the concerns being raised against AI generally can be tackled via categorisation.”
– Lee Kivell
The key thing is that a lot of the concerns being raised against AI generally can be tackled via categorisation. For example, there are often flags raised around “AI hallucinations”, and while this is a bad thing for decision making, it can be a good thing for content generation.
QA Financial: Finally, what testing challenges do you foresee in the next, let’s say, 12 months?
Lee Kivell: We have had a very heavy focus on improving the way we do testing over the last year, so most of the challenges identified have been resolved already, and the remainder will be addressed over the next 12 months. More generally though, there are the usual issues of keeping up with rapidly evolving technologies and ensuring that we have the appropriate skills and tooling in place to keep pace.
Lee Kivell spoke at the QA Financial Forum in London on September 11. More information can be found here.
TODAY IN LONDON
READ MORE
- Exclusive: Parasoft CEO talks competition, compliance and GenAI
- More code releases will lead to outages, warns Tricentis CEO
- Pentester Horizon3.ai shoots NodeZero platform into the cloud
- Sopra Banking Software wins BTK Bank deal for digital upgrade
- LambdaTest launches new real-time debugging tool
Become a QA Financial subscriber – for FREE
News and interviews * Receive our weekly newsletter * Get priority invitations to our Forum events
