Brian Dandoy can show his CFO exactly how Cast AI saves Paradox 40–70% on cloud costs. The graphs are impressive, and the leadership team loves them. However, what he can’t illustrate is the time his team gets back, now that they can focus on work that truly matters.
He can’t put on a graph the engineer who spent six months manually optimizing pods, only to have automation make that work obsolete in weeks. Or the 2am fires that no longer need to be put out. Or the 10 hours a week his team reclaimed from mindless CPU and memory tweaking.
“You can always make more money,” says Brian. “Time you cannot get back.’”
The benefits of Application Performance Automation (APA) are not always measurable, but they are real, and they compound. Brian’s role as the Infrastructure Architect Paradox involves managing teams that perform Kubernetes deployments. Part of that involves minimizing the time they spend on menial tasks so they can focus on the bigger picture.
To prove value to his managers, he goes beyond the traditional focus on cost metrics and looks at something more important: sanity, both his and that of his teams.
The six-month experiment that went nowhere
One of Brian’s team members spent a significant amount of time manually rightsizing pods. Six months later, when they checked with Cast AI, everything was already outdated. All that work didn’t even matter.
“Code changed,” says Brian. “We added more customers. Everything about your environment is constantly changing day-to-day.”
That’s the Sisyphean nature of manual optimization. Any time you make tangible progress, the tech changes, and you end up right back where you started. It’s like pushing a boulder up a hill only for it to slide all the way back down, over and over and over again.
It’s not just about wasted time. That time adds up to money…a lot of money. Engineers are skilled people, and they do not come cheaply. The average salary for a DevOps engineer typically ranges between $176,000 and $258,000 per year, depending on experience, location, and company size. When you factor in the constant need to optimize pods and the time they can spend on high-value work, it adds up to a significant amount of money being left on the table.
The triple value of APA: money, time, and sanity
When Brian presents to executives, he leads with the graphs. “Our number one thing that we used initially was just trying to show the number of our EC2 savings,” he says.
While cost savings can be illustrated, time and employee focus are transformational benefits that are harder to quantify but are crucial for success.
The visible: cost savings
The most dramatic and immediate change was the AWS compute reduction. Brian could put that on a month-over-month graph and put a big dollar sign on it.
Before Cast AI, Brian and his team were autoscaling groups with identical EC2 instances: “whenever we needed that one more pod, we spun up one more node.” This added up to a lot of wasted resources. Once Brian got the tangible, demonstrable results he needed to get executive buy-in, he could take things one step further.
The hidden: time reclaimed
Financial savings are easy to show leadership. The time saved for your DevOps engineering team is harder to visualize on a graph.
Developers often don’t know what resource requirements their applications need, and have to guess when asked.
“I can see they want to send me a shrug emoji back because they don’t really know,” Brian says. ”It’s really a guess.”
Cast AI helped Brian’s team eliminate a lot of manual, routine tasks:
- Manual right-sizing of daemon sets across different node types.
- Adjusting memory allocations for customer-specific workloads
- Responding to requests from devs who don’t know resource requirements
Brian could then spend that extra bandwidth on automation, building for 10x growth, and planning for partner ot client integrations.
Brian’s whole global team felt the impact. Engineers in different time zones weren’t getting pinged at 9pm for resource requirements.
The invaluable: peace of mind
When Brian saw that the infrastructure could run on its own, it took away a lot of unnecessary stress from his day-to-day.
“I’m no longer worried about things like ‘do I have the right memory settings?’” Brian says, “Is this pod going to run out of memory tomorrow? I just know that it’s being taken care of in the background.”
Brian’s team is distributed and uses asynchronous communication. Brian’s distributed team operates around the clock, and automation with APA has reduced the pressure of managing business operations 24/7.
He learned that he could trust the new system to run independently without extensive manual oversight.
“I can focus on other things,” he says, “Which is so valuable.”
Automation augments, it doesn’t replace
For all the undeniable benefits APA has provided him and his team, Brian maintains that the human element remains critical at every step of the way. There needs to be someone maintaining the automation so that it’s working properly and optimally.
Brian likens it to a guardrail. When he was managing the infrastructure himself, he was the guardrail. As his team scales with APA, the automation becomes the guardrail, and the humans on his team become the architects.
From human implementation to scaling, human oversight matters for many reasons.
For one thing, routine workflow audits will occur as AI agents become more commonplace. There’s important context that automation can’t grasp, too. Automation can run a workflow and do a task, but it doesn’t understand why it’s doing it. High-level strategic decisions still require human judgment to ensure they align with business goals.
“AI is there to help us, it’s not to replace us,” Brian affirms. “It’s not there to do my job for me, it’s to help me to do my job better and faster.”
Making the invisible visible to leadership
The time efficiency and better sleep schedule are all well and good, but how do you convince leadership they’re worth it?
Brian’s approach is to start with what executives understand: money.
“A lot of the work that infrastructure teams are doing is important work,” Brian explains. “It’s not the most glamorous type of work out there, but it’s very important. So being able to show things that will matter to people, whether it be a CEO or a CFO, of cost savings, performance improvements.”
One of the main challenges is in quantifying engineering time. It’s not just about saving on paying engineers for their time. It’s really about multiplying the impact of your engineers.
Time is the ultimate currency
APA isn’t only about cloud cost optimization. It’s about giving engineers their extremely limited and valuable professional lives back. That is the future of infrastructure work: humans do the human work, and automations do the drudgery and toil.
“You really have to make sure you’re finding the right places,” Brian concludes. “Finding things that either have a huge impact on our business, or maybe it’s something where I can do this really easily and get some easy wins to show the value of automation.”
Consider: Strategic APA implementation gave Brian’s team an additional 10 hours per week. What could your team accomplish with that time?
