Finance – AI + FinOps: from regulation to resilience – Business Reporter
Alex Gkiouros at DoiT examines how organisations can apply FinOps 3.0 principles to underpin a successful AI strategy
Last month, the next major compliance deadline under the EU AI Act came into effect. The legislation establishes guardrails for the development and deployment of AI, ensuring safety while fostering innovation. Imposing a risk-based classification system, the EU AI Act introduces different transparency obligations, depending on the AI system in question.
As such, businesses need robust systems for transparency and risk mitigation to avoid being fined for non-compliance. Across Europe, the new compliance deadlines send a clear message: adopting AI isn’t enough on its own – it also needs to be efficient, ethical, and financially accountable.
The European AI market is set to grow rapidly from $1.72 billion in 2025 to almost $11 billion by 2035, confirming that AI is here to stay. Yet, as regulatory pressure continues to build, many organisations risk repeating the same mistakes we saw with early cloud adoption: rushing to implement new technologies without any long-term strategy, only to end up with spiralling costs and business inefficiencies.
I’ll admit, it’s tempting to take a ‘tick box’ approach to compliance, meeting the bare minimum to pass an audit. It’s quick, measurable, and feels manageable, especially when a major deadline is looming. But this approach, unfortunately, only leaves organisations exposed to hidden costs later down the line.
So, how can businesses navigate this complex landscape? The answer lies in rethinking AI governance entirely, to see it not as a tedious compliance exercise, but as a framework that integrates compliance, financial oversight, and operational excellence.
Enter: FinOps.
Why FinOps is no longer just for cloud
At its core, FinOps is a financial operations framework, originally designed to enhance cloud efficiency. With visibility, collaboration, and accountability as its key tenets, FinOps principles promote a data-driven decision-making process. It also fosters financial responsibility, beyond the confines of finance teams.
The latest ‘iteration’ of FinOps, FinOps 3.0, takes things one step further. It moves beyond a cloud-focused mindset, from ‘business value drives cloud decisions’ to ‘business value drives all technology decisions’. As a result, teams broaden their focus beyond the cloud usage-cost equation to consider variables within a broader context. By introducing sustainability, unit economics, and real-time data insights, FinOps 3.0 embraces automation as a practical solution.
In other words, adopting AI with FinOps principles in mind grants financial visibility, alignment across teams, and cost control, all whilst being equipped for sudden compliance obligations.
Laying the groundwork for scalable AI
Now, with all of this in mind, let’s outline some practical ways to implement FinOps principles to AI adoption strategies using some real-life examples.
1 Align AI investment with business outcomes
The first step is connecting every AI initiative to measurable business objectives. Without this, teams end up with experimental spend that’s difficult to justify when budget reviews come around.
I’ve seen this play out repeatedly: engineering spins up LLM experiments across multiple projects, each team using different models and API keys, with no shared view of what’s actually being spent or delivered. One UK client we worked with couldn’t even tell which projects were consuming the most compute until we implemented cost attribution by team and engagement.
The fix isn’t complicated, but it does require discipline. Define what success looks like at the business-unit level – creative teams might measure output quality, engineering might track cycle-time reduction, operations might focus on workflow automation. Then tie spend directly to those outcomes using tools like DoiT’s Allocations, which map cloud and AI costs to internal teams or customer projects. That visibility turns vague “AI investment” into defensible ROI.
2 Use AI monitoring and automation
Manual oversight doesn’t scale. When you’re running AI workloads across multiple teams and providers, relying on someone to spot inefficiencies or compliance gaps in a spreadsheet is a recipe for missed issues and wasted spend.
The practical approach is continuous monitoring with automated response. Tools should flag anomalies in real-time, not just cost spikes, but usage patterns that suggest waste (like oversized models handling simple tasks) or risk (like API keys with overly broad access).
But detection alone isn’t enough. The next step is automation that actually does something, like pause a runaway API key, notify the owning team when a threshold is breached, or trigger a review workflow when a new model is deployed. This isn’t about removing humans from the loop; it’s about ensuring the right humans get involved at the right time, with the right context.
3 Generate transparency around AI costs and usage
Static dashboards showing last month’s bill aren’t enough when AI costs can spike overnight. A single misconfigured API key or an unexpectedly chatty LLM integration can blow through a quarterly budget in days.
The challenge is that AI spend is fragmented – OpenAI, Anthropic, Vertex AI, Bedrock — each with its own billing model and usage metrics. Without a unified view, teams can’t see which models are driving cost, which projects are underperforming relative to spend, or where there’s outright waste.
DoiT’s GenAI Intelligence addresses this by pulling cost and token usage across providers into one place. You can allocate by API key, model, or team, set anomaly detection to flag unexpected spikes, and automate guardrails (via CloudFlow) that pause keys or notify owners when thresholds are breached. The goal isn’t just visibility – it’s actionable intelligence that lets you scale, refine, or retire projects before costs spiral.
4 Promote cross-team collaboration
Bring finance, product, engineering, and compliance teams together to understand AI’s operational and financial impact. In other words, assign owners to specific AI initiatives to dedicated teams to promote responsible innovation. This can be reinforced with regular budget and resource utilisation reviews. The result? A culture of accountability through the project’s development lifecycle.
5 Future-proof all AI portfolios
Lastly, Governance that stops at “did we pass the audit?” is governance that’s already behind. The EU AI Act is just the beginning. Regulatory requirements will evolve, new compliance obligations will emerge, and the cost of retrofitting accountability into existing systems is always higher than building it in from the start.
Future-proofing means assessing AI investments across three dimensions: cost efficiency (are we getting value for spend?), compliance (can we demonstrate transparency and risk controls?), and sustainability (will this architecture scale without spiralling costs or technical debt?).
A FinOps culture creates the conditions for this. When teams have real-time visibility into AI costs, when anomalies trigger automated responses, and when cross-functional ownership is built into the operating model, you’re not just meeting today’s requirements; you’re building the muscle to adapt as the landscape shifts.
The organisations I see struggling aren’t the ones that moved slowly on AI. They’re the ones that moved fast without the financial and operational scaffolding to sustain it.
Alex Gkiouros is Cloud Architecting and Engineering Lead at DoiT
Main image courtesy of iStockPhoto.com and Just_Super
© 2025, Lyonsdown Limited. Business Reporter® is a registered trademark of Lyonsdown Ltd. VAT registration number: 830519543
source
This article was autogenerated from a news feed from CDO TIMES selected high quality news and research sources. There was no editorial review conducted beyond that by CDO TIMES staff. Need help with any of the topics in our articles? Schedule your free CDO TIMES Tech Navigator call today to stay ahead of the curve and gain insider advantages to propel your business!
Last month, the next major compliance deadline under the EU AI Act came into effect. The legislation establishes guardrails for the development and deployment of AI, ensuring safety while fostering innovation. Imposing a risk-based classification system, the EU AI Act introduces different transparency obligations, depending on the AI system in question.
As such, businesses need robust systems for transparency and risk mitigation to avoid being fined for non-compliance. Across Europe, the new compliance deadlines send a clear message: adopting AI isn’t enough on its own – it also needs to be efficient, ethical, and financially accountable.
The European AI market is set to grow rapidly from $1.72 billion in 2025 to almost $11 billion by 2035, confirming that AI is here to stay. Yet, as regulatory pressure continues to build, many organisations risk repeating the same mistakes we saw with early cloud adoption: rushing to implement new technologies without any long-term strategy, only to end up with spiralling costs and business inefficiencies.
I’ll admit, it’s tempting to take a ‘tick box’ approach to compliance, meeting the bare minimum to pass an audit. It’s quick, measurable, and feels manageable, especially when a major deadline is looming. But this approach, unfortunately, only leaves organisations exposed to hidden costs later down the line.
So, how can businesses navigate this complex landscape? The answer lies in rethinking AI governance entirely, to see it not as a tedious compliance exercise, but as a framework that integrates compliance, financial oversight, and operational excellence.
Enter: FinOps.
Why FinOps is no longer just for cloud
At its core, FinOps is a financial operations framework, originally designed to enhance cloud efficiency. With visibility, collaboration, and accountability as its key tenets, FinOps principles promote a data-driven decision-making process. It also fosters financial responsibility, beyond the confines of finance teams.
The latest ‘iteration’ of FinOps, FinOps 3.0, takes things one step further. It moves beyond a cloud-focused mindset, from ‘business value drives cloud decisions’ to ‘business value drives all technology decisions’. As a result, teams broaden their focus beyond the cloud usage-cost equation to consider variables within a broader context. By introducing sustainability, unit economics, and real-time data insights, FinOps 3.0 embraces automation as a practical solution.
In other words, adopting AI with FinOps principles in mind grants financial visibility, alignment across teams, and cost control, all whilst being equipped for sudden compliance obligations.
Laying the groundwork for scalable AI
Now, with all of this in mind, let’s outline some practical ways to implement FinOps principles to AI adoption strategies using some real-life examples.
1 Align AI investment with business outcomes
The first step is connecting every AI initiative to measurable business objectives. Without this, teams end up with experimental spend that’s difficult to justify when budget reviews come around.
I’ve seen this play out repeatedly: engineering spins up LLM experiments across multiple projects, each team using different models and API keys, with no shared view of what’s actually being spent or delivered. One UK client we worked with couldn’t even tell which projects were consuming the most compute until we implemented cost attribution by team and engagement.
The fix isn’t complicated, but it does require discipline. Define what success looks like at the business-unit level – creative teams might measure output quality, engineering might track cycle-time reduction, operations might focus on workflow automation. Then tie spend directly to those outcomes using tools like DoiT’s Allocations, which map cloud and AI costs to internal teams or customer projects. That visibility turns vague “AI investment” into defensible ROI.
2 Use AI monitoring and automation
Manual oversight doesn’t scale. When you’re running AI workloads across multiple teams and providers, relying on someone to spot inefficiencies or compliance gaps in a spreadsheet is a recipe for missed issues and wasted spend.
The practical approach is continuous monitoring with automated response. Tools should flag anomalies in real-time, not just cost spikes, but usage patterns that suggest waste (like oversized models handling simple tasks) or risk (like API keys with overly broad access).
But detection alone isn’t enough. The next step is automation that actually does something, like pause a runaway API key, notify the owning team when a threshold is breached, or trigger a review workflow when a new model is deployed. This isn’t about removing humans from the loop; it’s about ensuring the right humans get involved at the right time, with the right context.
3 Generate transparency around AI costs and usage
Static dashboards showing last month’s bill aren’t enough when AI costs can spike overnight. A single misconfigured API key or an unexpectedly chatty LLM integration can blow through a quarterly budget in days.
The challenge is that AI spend is fragmented – OpenAI, Anthropic, Vertex AI, Bedrock — each with its own billing model and usage metrics. Without a unified view, teams can’t see which models are driving cost, which projects are underperforming relative to spend, or where there’s outright waste.
DoiT’s GenAI Intelligence addresses this by pulling cost and token usage across providers into one place. You can allocate by API key, model, or team, set anomaly detection to flag unexpected spikes, and automate guardrails (via CloudFlow) that pause keys or notify owners when thresholds are breached. The goal isn’t just visibility – it’s actionable intelligence that lets you scale, refine, or retire projects before costs spiral.
4 Promote cross-team collaboration
Bring finance, product, engineering, and compliance teams together to understand AI’s operational and financial impact. In other words, assign owners to specific AI initiatives to dedicated teams to promote responsible innovation. This can be reinforced with regular budget and resource utilisation reviews. The result? A culture of accountability through the project’s development lifecycle.
5 Future-proof all AI portfolios
Lastly, Governance that stops at “did we pass the audit?” is governance that’s already behind. The EU AI Act is just the beginning. Regulatory requirements will evolve, new compliance obligations will emerge, and the cost of retrofitting accountability into existing systems is always higher than building it in from the start.
Future-proofing means assessing AI investments across three dimensions: cost efficiency (are we getting value for spend?), compliance (can we demonstrate transparency and risk controls?), and sustainability (will this architecture scale without spiralling costs or technical debt?).
A FinOps culture creates the conditions for this. When teams have real-time visibility into AI costs, when anomalies trigger automated responses, and when cross-functional ownership is built into the operating model, you’re not just meeting today’s requirements; you’re building the muscle to adapt as the landscape shifts.
The organisations I see struggling aren’t the ones that moved slowly on AI. They’re the ones that moved fast without the financial and operational scaffolding to sustain it.
Alex Gkiouros is Cloud Architecting and Engineering Lead at DoiT
Main image courtesy of iStockPhoto.com and Just_Super
© 2025, Lyonsdown Limited. Business Reporter® is a registered trademark of Lyonsdown Ltd. VAT registration number: 830519543
source
This article was autogenerated from a news feed from CDO TIMES selected high quality news and research sources. There was no editorial review conducted beyond that by CDO TIMES staff. Need help with any of the topics in our articles? Schedule your free CDO TIMES Tech Navigator call today to stay ahead of the curve and gain insider advantages to propel your business!
