Experts share 7 real-world FinOps mistakes to avoid

In theory, innovations in cloud cost management tools provide enterprise leaders with the knowledge needed to drive efficient cloud usage. Increased visibility into cloud spend details should help everyone get on the same page, realize savings and align business growth with the right cloud infrastructure.   

But the real world is full of messy details, cultural pushback and other sources of friction for grand visions of cloud and AI spending nirvana. Enterprise leaders often discover numerous things they wish they had done differently to streamline their FinOps journey. This could include cultural resistance, rushing to implement without proper planning or tool issues.  

Below are seven critical mistakes to avoid when starting your FinOps journey, such as underestimating AI, mishandling the roll out and not considering the cultural impact. 

1. Don't underestimate your AI footprint 

The growing hype around AI is pushing many enterprises to drive AI across their organizations. The cloud is a natural proving ground for many of these AI experiments. This can come with even less visibility and control than traditional cloud apps.  

Will Thomas, managing director and cloud optimization lead at Protiviti, explained, "While organizations are getting their data classification and governance aligned to compliance objectives and also defining their AI use cases, solutions and offerings, there should be a strong consideration in understanding the unit economics of enabling AI.”  

Thomas said many of the companies he is working with wish they had spent more time understanding the costs of AI workloads and their predictability. This requires improving visibility into AI chip usage and storage costs. FinOps for AI emerged as the top forward-looking priority, according to the State of FinOps 2026 Report by the FinOps Foundation. 

This kind of visibility can help reevaluate reservations to optimize the use of specialized resources. Predictive models and scenario planning can help anticipate fluctuations. It can also guide the development of additional guardrails for budget thresholds and automated alerts.  

"As AI is the current big thing in tech, the new challenge for FinOps practitioners is defining unit economics of cloud spend for AI solutions by breaking down costs into measurable units tied to business value," said Thomas. 

2. Don't treat FinOps as a one-off project 

Another problem that can emerge is implementing a FinOps project as a one-off cost-cutting program to sprint ahead. Himanshu Jain, a partner in Kearney's Digital and Analytics practice, said enterprise leaders wish they had treated FinOps as a continuous loop. This requires thinking about how to build capabilities to inform users, optimize resources, and fine-tune operations to support continuous improvement.   

A good practice is to find ways to meet engineers where they are and put cost signals into developer tooling, including dashboards and development tools. This shortens the loop between insight and action. It's also important to centralize telemetry and decentralize actions so that patterns identified in a Center of Excellence (CoE) can be paired with team-level accountability for optimization. Also consider ways to move from a total-bill approach to matching costs per unit of value, such as requests, orders and API calls. This can also improve forecasting even when savings are not the primary goal.  

3. Don't mishandle the rollout 

Tatum Tummins, senior product manager at Kion, a FinOps platform, observed that many organizations struggle with FinOps not because the practice itself is flawed, but because the rollout is mishandled. For example, sometimes overemphasizing the term FinOps can trigger resistance, particularly from teams that have had bad experiences with CCoEs or anticipate increased bureaucracy.  

Tummins recommends teams focus on the work rather than FinOps as a branding. This might mean quietly implementing FinOps-aligned processes, such as business reviews, cross-functional conversations and sharing early wins, to promote trust and experimentation. Only call it FinOps once these early efforts are successful,  

4. Don't underestimating the cultural element 

FinOps fails when it's seen as a policing function instead of a partnership.

Other FinOps projects run into pushback when leaders fail to appreciate the cultural element. "FinOps fails when it's seen as a policing function instead of a partnership," said Tummins.  

He recommends developing "people pillars" composed of a small set of champions across engineering, finance and product departments. Part of this process requires learning how to support these pioneers, prioritizing early wins and then helping them to share their experience with other teams across the organization.  

"By helping these individuals succeed in their own work and openly celebrating those successes, FinOps became something people wanted to participate in rather than something imposed on them," said Tummins. This can also help create a shared understanding across different departments and groups.  

5. Don't treat FinOps like a software installation 

Enterprises sometimes treat FinOps like a software installation rather than an operating model and cultural shift. This can lead to disenchantment with the lackluster results.   

People assume that once they turn on the new tool, they will immediately save 20 to 30%. But in reality, pure visibility delivers 3 to 8% savings in the first year for almost everyone.

"People assume that once they turn on the new tool, they will immediately save 20 to 30%. But in reality, pure visibility delivers 3 to 8% savings in the first year for almost everyone," said Max Printz, principal customer success manager at Harness, a DevOps platform. 

A jump to higher levels of savings typically comes only when teams can layer in processes for automated rightsizing, intelligent commitment purchasing and auto-parking idle environments. "Visibility is the entry ticket, automation is what actually pays the bills," said Printz. 

6. Don't create dashboard friction for developers 

Another mistake lies in failing to find ways to build motivation for cost savings across the engineering team. As a result, many enterprises create a new dashboard and hope for the best. "However, engineers treat cost dashboards exactly like they treat security or tech-debt dashboards; they ignore them unless cost becomes part of the development workflow," said Printz. 

A better practice is to surface cost management directly within the tools developers are already using for CI/CD checks, pull request management and golden path templates. Removing friction and enforcing best practices tends to work better than adding another dashboard.  

7. Don't ignore tagging 

Tagging is the process of adding standard labels to cloud resources to make it easier to account for spend. It plays an important role in various FinOps processes, and enterprises sometimes rush through a tagging implementation to get the ball rolling.  

Printz said that it's a mistake to think that tagging can be cleaned up later. Untagged or poorly tagged resources can account for significant spend, particularly in large environments. "If organizations don't solve tagging governance in the first 90 days, their chargeback or showback efforts collapse, and the entire FinOps program grinds to a halt," he said. "It's still the single biggest blocker we see." 

George Lawton is a journalist based in London. Over the last 30 years, he has written more than 3,000 stories about computers, communications, knowledge management, business, health and other areas that interest him.