IIA clients leverage analytics maturity assessments, battle-tested frameworks, and cross-industry collaboration to prioritize and execute strategic enterprise data and analytics initiatives. We regularly check the pulse of trending topics for our community and facilitate critical conversations in virtual roundtable format for peer-to-peer exchange.
In a recent IIA roundtable discussion, data and analytics leaders from diverse industries explored strategic approaches to forecasting and budgeting for AI investments in 2025. They shared candid insights into the challenges of forecasting costs, measuring returns, and aligning funding strategies with business priorities. The discussion highlighted both shared struggles and strategies that can serve as valuable lessons for organizations navigating similar terrain.
Here are the key takeaways:
1. The Complexity of Forecasting and Allocating Cloud Costs
AI workloads bring inherent unpredictability in cloud costs, making accurate forecasting a significant challenge for many organizations. Factors such as compute demand spikes, data transfer costs, and storage needs often catch enterprises off guard, especially when transitioning to cloud or multi-cloud environments. One government representative highlighted the difficulty of projecting costs for AI models, especially in a fixed-budget environment where flexibility is limited.
Participants shared strategies to navigate this challenge, such as using reserved instances for compute to secure significant cost savings. However, they cautioned that this requires precise forecasting to avoid waste. Another approach involved setting up cost codes and detailed time tracking to increase visibility into project-level spending. Yet, even with these measures, leaders agreed that the dynamic nature of AI initiatives—such as sudden shifts in use cases or unanticipated events like COVID—makes complete cost predictability elusive.
2. Balancing Centralized and Decentralized Budgeting Models
Organizations face a critical decision when determining how to structure AI budgets. Centralized budgeting, where all AI investments are managed by a single team, offers greater control but can blur visibility into specific AI-related expenses. Conversely, decentralizing budgets to business units can enable more targeted investments but risks creating silos and misaligned priorities.
Leaders discussed hybrid models as a potential solution. One organization adopted quarterly re-forecasting and established investment pools for strategic projects, enabling flexibility without sacrificing oversight. Another piloted a collaborative budgeting model where business partners and analytics teams jointly prioritized initiatives. While promising, this approach requires significant cultural change, as it demands cross-functional alignment and trust. Leaders agreed that whichever model is chosen, balancing flexibility with strategic alignment is key.
3. The Elusive ROI of AI Investments
Measuring the return on AI investments remains a perennial challenge. While many organizations know AI adds value, quantifying that value in dollar terms often proves difficult. For example, participants described how AI tools like internal ChatGPT save time and increase productivity, but those gains don’t directly translate into measurable financial outcomes unless tied to workforce reduction or revenue growth.
Some leaders shared creative approaches to demonstrate ROI. One organization ran controlled experiments to measure time savings achieved by AI tools, using these metrics to approximate productivity gains. Another company tied AI to operational efficiencies on its manufacturing floor, such as reducing training times from months to hours. While imperfect, these methods provide a starting point for justifying investments. However, participants noted that bridging the gap between productivity metrics and tangible financial impact remains a significant hurdle.
4. Innovative Funding Mechanisms for AI Projects
To foster innovation, several organizations are experimenting with creative funding models. Hackathons and “Shark Tank”-style competitions emerged as popular approaches, offering teams the opportunity to pitch ideas to executives and secure funding for promising projects. This not only incentivizes employees but also allows organizations to crowdsource ideas for AI use cases.
One participant described a model where employees pitched AI projects to an executive panel and, if successful, received funding and continued involvement in the project’s development. However, leaders cautioned that follow-through is essential; without a clear process for scaling successful projects, initial enthusiasm can falter. Another approach involved reserving small budgets for proof-of-concept (POC) projects, enabling experimentation without significant risk. While this strategy allows for early-stage testing, it also requires a robust mechanism for scaling pilots that demonstrate value.
5. Integrating Business Partners into Budgeting Processes
Aligning AI budgets with business priorities is critical, and many participants highlighted the growing role of business partners in shaping funding decisions. Collaborative models, where analytics teams and business units jointly determine priorities, can help ensure alignment and maximize enterprise-wide impact.
One organization shared its experience piloting such a model, where product owners worked closely with business leaders to co-develop funding proposals. This approach improved alignment and allowed for more targeted investments in high-priority initiatives. However, leaders noted that building these partnerships requires consistent communication and trust, as well as a willingness to adapt budgeting processes to accommodate diverse perspectives.
IIA virtual roundtables are exclusive, invite-only discussions designed to promote peer-to-peer exchange on pressing challenges in the data and analytics community. Seats are limited and reserved for C-suite data and analytics leaders or equivalent at mid- to large-sized enterprises. Conversations are geared toward non-digital native companies. If you meet these criteria, contact us for more information.