A conversation with Holly Craig, Lead Software Engineer at WiseTech Sydney.
At WiseTech Global, AI has changed how our product teams build, test, learn and improve. For software engineers, that means using AI to move faster toward their end goal: understanding unfamiliar problems, generating, and validating ideas, automating repeatable steps, and applying human expertise where it matters most.
We spoke with Holly Craig, a Lead Software Engineer in our Sydney office, about the role of AI in her day-to-day work, how she decides what to automate, and what her team has learned from putting agents into real engineering workflows.
How AI is changing daily engineering work
The question for Holly and her team moved some time ago from whether AI can help, to how to use it well: when to let it accelerate the work, when to challenge it, and where engineering judgement remains essential.
AI gives Holly the confidence to move into complex work faster, especially when she needs to understand highly technical documentation, unfamiliar codebases, or a new problem domain.
“AI has lowered the barrier to entry for me to work on tasks that may have, in the past, intimidated me,” Holly says. “Information is now at my fingertips, so I’m less blocked by highly technical documentation or unfamiliar codebases. If I don’t understand something, I can keep asking AI questions until it clicks.”
From quick starts to better judgment
Holly uses AI to reduce manual effort and speed up discovery, then applies her own judgment to shape the solution, test assumptions and focus on the design and problem-solving work that creates the most value.
"AI has completely eradicated the tedious and repetitive parts of my job," Holly says, "allowing me to spend more time on the design and problem-solving phases of a task."
One example is a weekly team report that used to be pulled together manually. Now, it is automated and lands in a group chat 10 minutes before the meeting starts, giving the team the information they need.
"If it's repetitive, tedious and structured, it's usually a good candidate for automation," Holly says. "AI agents now help with much of the defect review and investigation work that used to take up significant developer time. That means I can spend less of my day fixing issues and more time designing and building new features for customers."
Find the root cause, and solve for that
One of WiseTech’s core mantras is to avoid quick fixes and solve the root cause. Holly built an AI agent that helps engineers reach that point faster and with less manual effort.
Engineers used to spend hours tracking down obscure defects. Now, an agent has already investigated the defect by the time it arrives in a work item ready for an engineer to review.
The Agent gathers evidence from across the stack - Elasticsearch logs, Sentry stack traces, MongoDB data integrity checks, and recent GitHub changes - to build a picture of what's behind the issue. Alongside the evidence, it provides reproduction steps, root-cause analysis, and a recommended code fix.
Holly's role shifts from investigating the problem to exercising engineering judgment: evaluating the evidence, challenging the analysis, and applying her expertise to decide whether the agent's solution is the right one.
“Even if the analysis is wrong, we still have all the evidence at our fingertips and can see how the agent arrived at its conclusion, so it still saves us a significant amount of time,” Holly says. “Previously, we'd have to gather all of this by hand, which took up a large part of my time.”
Building confidence in regression testing
Regression testing is another of the most important yet time-consuming parts of software delivery. Holly built an agent to help with that too.
The regression testing agent can simulate slower network conditions, flag console errors, and recreate issues that only appear under specific timing or system conditions.
By bringing test planning, execution, and reporting into one flow, it helps teams reduce manual effort, improve quality, and release with greater confidence. The ripple effect means that customers benefit from more product enhancements, delivered sooner.
Learning through conversation
Holly built her AI capability as much from other people's experiments as her own. Every engineer develops different shortcuts, tools and ways of working that Holly uses as inspiration for her own team.
“I love asking other people about their workflows and optimizations and stealing the parts that would work for me and my team,” Holly says.
Holly's team runs a fortnightly AI roundtable, where people share what they have tried, what they have learned, and what others might reuse. These conversations ensure that practical ideas spread across teams, turning individual experiments into shared ways of working.
Looking ahead
As AI accelerates work across WiseTech and pushes repetitive tasks into the background, Holly's next frontier isn't freight, logistics or software. It's life admin.
"I wish I never had to fill out a form, write an email or plan an event ever again," she says. It sounds familiar, because it is.
The same instinct that drives Holly's engineering work, letting AI take the tedious part, so people can spend their time on the parts that need real judgment, is one she'd happily apply everywhere in her life, not just at work.
As AI removes more of the repetitive load, engineers like Holly can focus on what only they can do: understanding complex customer scenarios, navigating technical trade-offs, applying deep engineering expertise, and finding even better ways to solve the challenges that matter most to our customers.