You’re the CIO at ABC Inc., a fictional rapidly scaling tech startup. In 2023, you implemented a new cloud finops system to streamline the management of its extensive cloud infrastructures. In 2024, the system unveiled several previously unknown technical debts. You’ve been summoned to a meeting with the CEO to address these issues.
The finops system has determined that this technical debt has led to more than $20 million in lost revenue to the business. This includes general inefficiencies, such as overprovisioning public cloud storage systems, as well as more strategic problems, such as the need for the system to provide the agility and scale to support moving into new markets and other innovations.
When finops AI turns on you
As the CIO, you were previously praised for cost-effectively sustaining the company’s rapid growth. Now you’re at the center of criticism based on a report that came from a cloud finops system that you paid for and implemented.
Those who reviewed the AI-enabled finops reports claim that you prioritized short-term gains over long-term efficiency, resulting in underoptimized cloud operations and bloated costs. Criticisms highlight inefficient resource allocation, excessive idle instances, inflated data storage costs, and a lack of automated processes.
The most damning: Several market opportunities and acquisitions were likely missed due to this short-sightedness, which is being put on you, the CIO. The necessary cloud technology to provide the agility and scalability required to access those prospects was not there.
It once was harder to define values, but now the finops system can describe them to the dollar amount. Thanks to the intermixing of large language models (LLMs), you can see the growth patterns of similar companies aligned with their IT solutions and compare them to the cloud solutions you designed and implemented.
Does this scenario sound stressful? I suspect we’ll see similar cases play out in 2024 and 2025. I’m not sure that today’s finops systems are there yet with AI, nor are they trusted enough to create data that would cause such an issue.
However, we’re getting close to the point where finops, cloudops, secops, and other x-ops will provide productive information by leveraging AI to not only find issues but also to look for benchmark performance relative to other organizations. For those of you who have been through audits that include this type of analysis, this will be much the same, but instead of coming from an outside consulting firm with humans, a set of AI-driven processes will do it. Imagine being audited daily.
Upsides and downsides
Most businesses can now afford cloud finops, but they may not be prepared for the good and bad aspects of this kind of data being available to any company.
The downside is shown by our fictional case study. There will be situations where the very finops system that a CIO or other IT leader deployed becomes their accuser. You can look at the company’s culture as to how they will likely handle this, but the board of directors and other executives have a duty to the shareholders to do something, meaning that this kind of stuff could get you fired.
Of course, much of it will roll downhill and the decision-makers will pass the buck to the line managers in many instances. That’s not entirely unfair, considering that most IT executives rely on recommendations from their reports. However, the final decision is their responsibility.
The upside is that I can stop having silly arguments about the choice and configuration of cloud technology. Finops systems can look at the value of each tech configuration and usage, new or old, checking net-new cloud architecture as well as the past decisions of others. It’s much more valuable to see this data ahead of implementation, rather than making a multimillion-dollar mistake and having to fix it later. That’s really what technical debt is.
Even the more strategic values that are now impossible to define can be put on the table, such as the value of growth and agility. Try going down the “strategic value of cloud computing” rabbit hole when doing cloud architecture. Still, it’s more important to get strategy correct than the tactical cost-efficiencies everyone seems obsessed with, considering the much larger impact.
This too can be managed
First, I’m not a fan of being punitive around these types of mistakes. Most people are well-intended and make decisions based on what they think is right. This includes IT executives, consultants, and technology providers.
Of course, biases can cause many of these negative outcomes. Some people will only implement a single public cloud brand, and others avoid the public cloud altogether. Any extreme is almost always wrong and leads to the types of negative outcomes I’ve mentioned. This includes those who try to avoid risk by doing nothing at all. Inaction is perhaps the worst thing you can do, and finops systems will uncover that as well.
Although AI will make finops tools much better and useful, they should be considered as sources of information and not be allowed to make decisions. I suspect that many will find this new information disruptive, considering that it’s judging current and past decisions made by humans.
We must find a productive way to leverage finops to make us better at using cloud computing. That’s the bottom line.