Indian enterprises are expanding their SaaS portfolios, relying more on contractors and stepping up AI experimentation. Yet visibility into these costs remains limited, with spending scattered across business units or tied to legacy contracts
According to Vineet Sawant, author of The CIO’s Guide to Enterprise FinOps, CIOs may alter this by tapping beyond cloud-only FinOps. In an exclusive interaction with CIONOW, he discusses on how tech leaders can identify waste across the entire IT portfolio, set consumption accountability, and free up up to 30% of spending to invest in business innovation, modernization, and artificial intelligence.
Traditionally, FinOps was looked as a control variable for public cloud costs, but these days, public cloud accounts for less than 25% of the usual company tech budget. Why is it dangerous to focus only on the cloud?
Cloud FinOps is a subset of enterprise FinOps. Whilst this is often associated with AWS and Azure bills, it’s actually under the control of the CIO (infrastructure, software, services, contractors and labor).
In cloud FinOps, you will often see the most obvious areas as being your workloads, AI token usage and overprovisioning. However, the real pain points are in license waste, duplicated SaaS tools, software, shadow IT, long standing contractors and fragmented vendor contracts. Software is often a good place for CIO’s looking for quick wins as it can deliver savings without any big investment into new technology and processes.
What are the most common hidden costs that enterprises eventually overlook?
Cluttered data is one of them. Even if a user spends $20 a month for a tool, it becomes a significant line item when you increase that amount over thousands of employees.
There are also people who are over-provisioned on premium software editions when a basic edition would meet their needs. Shadow IT is another major issue, especially with AI tools, where employees may sign up for individual subscriptions that slip under normal audit thresholds.
Contractors are another area. I have seen contractors remain with companies for 10 years after starting as a three-month engagement. If someone is that valuable, perhaps they should be converted into an employee rather than continuing at a contractor rate that may be twice as expensive.
Aside from upfront costs, what should CIOs consider when looking to compare cloud infrastructure with an enterprise data center?
The data center will be an asset that is depreciating but there will be ongoing opex costs associated, including power, cooling, maintenance, staff, licenses, and infrastructure costs. Cloud is flexible and less capital intensive. However, if you’re not managing your consumption well, you can end up paying more.
CIOs need to consider capital and operating costs. A hybrid solution can often get the best outcome. It can be crucial to have control over data, latency, security, and consistent consumption. You will gain scalability and agility by using the public cloud. It all boils down to what your task demands and what is most economical.
Generative AI means organizations now need to think about new costs associated with tokens. How can organizations avoid AI costs spiraling?
One of the main issues with generative AI is consumption. You may have a 30-user pilot, where consumption seems low. Once you scale your application to 30,000 users, the economics can change. It is important to understand the costs around AI. How much will you be consuming, how many tokens, what models are you using, and how much will the application cost you to scale?
A good option is to use Open-Source Large Language Models or bespoke AI models. Using these within your enterprise data centers on GPU infrastructure such as NVIDIA AI computing will reduce reliance on token costs. You will also have more control over data, latency and compliance. However, you will still have hardware, maintenance, power and operating costs. CIOs will need to consider these factors.
Should FinOps be focused on reducing token consumption rather than understanding the value AI applications are providing?
It’s not about reducing token consumption. You need to understand what value you’re getting for your money. If your AI application is providing value in productivity, efficiency, or revenue, then higher token use may be justified. It comes down to whether you are getting value from the spend.
CIOs need to consider if the model you have chosen is appropriate. Start using lower cost models and only move to high-cost models when you have a business requirement. Getting FinOps involved early on will help. Get them involved in the pilot and planning process. This will allow you to understand what the consumption will be, have controls in place and understand the costs. As AI gets adopted, you will want to understand token consumption.
While infrastructure can be a large part of IT spend, vendors and staff costs can also take up a big chunk of the technology budget. Where should CIOs look to optimize?
Vendors and staff can be looked at in the same way. Look to reduce vendors. If you have multiple vendors providing the same things, it can create a lot of overhead. Consolidating vendors where possible will help give you a better negotiation position.
Consider SLA’s. Sometimes businesses pay for 24×7 support but don’t really need it. If you can run the service with 18×7 support, then consider whether you need to pay for a high SLA. Consider staff. Are you using external contractors? Are you paying for deliverables or the number of people?
Sometimes you will have contractors that have been in place for a long time and costs can spiral. This can result in ‘headcount fraud’. Monitoring overtime and location of staff can also provide benefits. Overall, you should regularly review your recurring vendor and staff costs.
Software licenses and long-term vendor contracts often have auto-renewal. What should CIOs consider when looking to renegotiate?
With software licenses, check if you are paying for premium versions when you don’t need them. Ensure you are not paying for the same number of licenses as last time; understand how many people are actually using it. Consider payment options. Can you have tiered contracts? Are there any commitments that can be deferred? Can you align payment to consumption? Don’t assume you need to pay for the same amount as last year. Regularly review your vendors to understand if you are paying for anything that isn’t justified.
With business teams buying their own SaaS and playing with AI it can be difficult to manage shadow IT. How can CIOs manage this without limiting business agility?
The challenge is to ensure you are getting visibility without limiting procurement. Often business teams will buy applications to solve an issue. With the increase in SaaS and AI, this has become more common. However, when multiple departments buy similar applications, you can end up paying multiple times. You may also have licenses that aren’t being used.
A good option is to look at license harvesting. Before you go to purchase more licenses for a new requirement, look to see if you have any licenses that can be reclaimed. You don’t need to invest in a costly solution to start with. A simple spreadsheet to track what software you own, licenses, consumption, and renewal dates can show you where you are paying unnecessarily.
As your organization matures, you could put this into an automation solution using ITSM tools. Ultimately, you want to get visibility of software consumption but give business units flexibility.
How can organizations ensure they don’t create another bureaucracy by optimizing software?
The aim should be to gain visibility. CIOs need to understand what applications are being used, who owns them, whether they are duplicating anything, and whether we are getting value from the licenses. Regularly review software to understand if there are any modules that aren’t being used. Are there any duplicates and licenses that can be reallocated?
Ensure your procurement policy allows business teams to easily request new applications. If it’s too complicated, they wll find alternative ways to get what they need. Good governance will make technology more sustainable.
You advocate using the Run/Transform/Innovate 360 framework to manage technology spend. How does this framework help CIOs approach their IT budget?
The framework helps break down technology spend into three categories. First is the cost to run your business. Often this will be around 70% of your IT budget. Second is transform, where you are investing in modernizing your infrastructure and applications. This could be to reduce technical debt, improve security, and increase integration. And last is Innovate, where you are spending on new capabilities.
The issue with this is that a large chunk of your budget goes into Run, and you don’t have much flexibility to invest in Transform & Innovate. This is where FinOps can help. By identifying waste, reducing software licenses, optimizing infrastructure, and improving vendor contracts you can identify savings in your Run spend. This can then be reinvested into modernizing your organization.
The framework also helps with communicating with business leaders. Rather than just telling them that you have infrastructure, software and operating costs, you can tell them how much you are spending to keep your business running, how much you are investing to transform and how much you are investing for growth.
What impact will this have in the long term if organizations continue to invest in Run costs?
You will incur technical debt if you consistently spend the majority of your tech budget on maintaining your systems. Infrastructure becomes less adaptable, security becomes more complicated, integrating legacy apps with new platforms becomes challenging, and expenses can skyrocket.
Eventually, you will get to a point where it becomes more expensive and difficult to modernize. CIOs need to continually review the balance between Run, Transform and Innovate. You don’t want to stop running costs; businesses need reliable systems and infrastructure. However, you need to ensure they are being managed efficiently so they don’t eat up your whole tech budget. Having a structured FinOps approach can help.
With AI, hybrid infrastructure, and decentralized technology adoption, how should CIOs approach FinOps going forward?
FinOps needs to be viewed as a wider financial optimization. You can’t just look at your cloud consumption and identify unused VM’s. You need to have visibility across all of your technology spend. Including physical infrastructure, software, vendors, staff and AI.
As AI is becoming more prominent, it’s becoming more important. Traditional models are changing. Token costs, GPU infrastructure, and increased compute requirements all add new considerations. However, the basics still apply. You want visibility, accountability, and to understand how your investment is providing value.
The CIO’s role is to ensure that your technology spend is being spent strategically. Ultimately, having a good FinOps approach will mean your current technology investments are working harder for you. Giving you flexibility and investing in your next phase of growth.
