Trusted by some of the world’s largest banks to securely process and settle billions of dollars of cash and securities every day is a privilege and responsibility the team here at Baton take very seriously. Developing and deploying operationally resilient solutions is fundamental for our business, so we take a rigorous approach to quality assurance (QA) testing.
It’s crucial that the solutions we deploy to clients meet (and ideally exceed) their expectations both in terms of quality and performance. As such, functional testing cycles continue until the desired quality standards are consistently met and, as is often the case with software development, this can prove very time-consuming.
“Generative AI has freed up more time to focus on identifying and working on more complex test cases”
Keen to deploy diligently tested solutions out to clients faster, our testing team, which I’m part of, has been looking at how we can increase the productivity and efficiency of our functional testing process. This led us to exploring how the latest advancements in generative artificial intelligence (AI) could prove advantageous by providing an automated means of writing robust and efficient test scripts.
“..the team is now able to be more innovative and productive”
Looking for a solution that would work well within our existing architecture and workflows, and enable our team to achieve maximum test coverage (including all edge cases) while accelerating the overall functional testing process led us to considering Postman’s recently introduced generative AI assistant, Postbot.
Postman is a collaborative platform for API development and our go-to tool as a business for functional API testing. Postbot, which when we first started to explore it was still in beta, is capable of generating test scripts from simple, yet precise queries in English.
A large proportion of the functional tests executed for our solutions are for an extensive set of basic test cases. For example, we might want to check that a collateral movement instruction will be executed on a set date, in this instance we might want to give Postbot a query to check that the system knows the trade date must be prior to the value date.
Generating basic test scripts was previously a relatively manual step, however using Postbot as a team we’ve been able to automate the generation of test scripts and then run the test at the click of a button. Freeing up more time to focus on identifying and working on more complex test cases the team is now able to be more innovative and productive. In fact, using Postbot has actually increased the test coverage we can achieve.
“Using Postbot has actually increased the test coverage we can achieve”
For us, it’s not about using AI to replace the wonderful humans working here at Baton, but providing our colleagues with the tools to work more effectively and achieve more at an accelerated rate. All of which ultimately benefits our customers because they get an even more rigorously tested solution deployed in an expedited time-frame.
“Using AI for us is about providing our colleagues with the tools to work more effectively and achieve more at an accelerated rate”
By exploring how we could use generative AI we’ve also identified ways in which Postbot could be enhanced to increase capacity. One of the benefits of working with tools like this in the Beta stage is that we’ve been able to provide this feedback to Postbot’s developers, which will hopefully assist them in building a tool that will aid the industry on its AI journey.
The use of generative AI for functional testing is still in its early stages, but as a business we’re witnessing the transformative power that comes from being able to use these tools to accelerate existing workflows. What we’ve discovered is that AI is not only allowing us to deploy a higher quality solution faster, but that it also presents a means by which we can effectively scale our deployment efforts as we onboard more clients and the business continues to grow.
“AI is allowing us to deploy a higher quality solution faster and presents a means by which we can effectively scale our deployment efforts as we onboard more clients and the business continues to grow”