Michael Bayer is CFO of Wasabi Technologies, a cloud storage startup that reached unicorn status—valuation of over $1 billion—last year. When Bayer joined the company in 2018, he was the finance team, and handled most processes manually. Five years on, he oversees a business enablement team, including finance and accounting staff of ~30 people and has implemented a best-in-class tech stack. He spoke with CFO Brew about finance automation, obtaining staff buy-in around technology, and the risk and promise of generative AI.
This interview has been lightly edited for length and clarity.
How has the company’s tech stack changed since you started?
We’ve introduced automation really everywhere we could. We’ve implemented Salesforce for CPQ. We’ve put in Tableau on top of Snowflake which runs on top of Wasabi storage. So now we’re in the process of linking all that together. And it’s more of a process than a project because the technology continues to evolve.
What factors do you look for when choosing software to add to your tech stack?
We’re scaling. We’ve slowed down to more than 50% growth a year, as we’ve gotten bigger, but it’s still hyper growth. We’re always picking systems which have the ability to scale to an order of magnitude bigger than what we are today, because we will be there by the time the system is in place. And implementation will get incrementally harder the longer we wait.
How do you decide how much automation is appropriate?
I don’t want to implement 100% automation because that makes things very, very difficult to change. Yes, you design to be able to manage exceptions. But if there are too many exceptions, and you have too manual intensive a process, you’re going to be error-prone and it’s just not going to work for scale. And if you have too few exceptions, then you’re not able to flex, to be dynamic as the business changes. And so we’re trying to find that balance and build systems that have a little bit of flexibility, but automate the vast majority of transactions.
How do you ensure your staff is comfortable with the technology?
As a CFO, you really have to think about how you are going to balance automation with the demands of your workforce. You can’t tell people “Hey, help me with this automation project so you can work yourself out of a job.” It’s important to show people, “Look, we’re growing very quickly, and we’re creating opportunities for career progression because you can automate pieces of your…role that are repetitive, which enables you to take on new challenges.”
News built for finance pros
CFO Brew helps finance pros navigate their roles with insights into risk management, compliance, and strategy through our newsletter, virtual events, and digital guides.
How do you see generative AI changing the finance function?
I heard of a CFO who had a reconciling challenge between their payment receipts and their invoices every month. There was a data cleanup that they had to do manually. They told one of the [generative AI] tools, “Write me a Python program that does this” and it gave them a Python program that did that! They asked, “How do I run a Python program?” and they got instructions on how you upload Anaconda and get it running. And then with a little bit of tweaking, and some Python expertise applied, bang, they had their approach.
These tools aren’t just about automation; they’re about changing our interaction with systems and have us thinking differently about how we accomplish tasks. And I’m a firm believer that the more automation we can apply, the more intellect we can apply. We can spend more of our time thinking than doing and that’s what we’re all pursuing.
Are you experimenting with generative AI at all? What are the risks?
We’re exploring what we could do with it. But today, there are very few enterprise class tools that I can deploy [and] know that my data is safe, know that it’s not being used to teach someone else’s model to do something. I think about error rates—the systems are not 100%, so there’s still some level of manual interaction that’s in them. And then I have to think about how much work I want to put into applying one of these models, when, literally next week, something may come out that will obviate everything that I just did.
How can you find niche systems that you trust? I’m going to underscore that: that you trust. Am I going to unleash an AI that’s got this automated collections engine that starts dunning my customers too often, just because it sees a pattern that we didn’t think about? That’s where implementation is a little risky.
I’m excited about what’s coming. I can’t grow headcount 50%–60% a year forever. So I want to equip the team with tools to get some of the repetitive tasks out of the way. And if those are AI tools, great. If they’re just simple automation, great.