Channel: IBM
Date Published: 2024-12-19
Summary
This episode of AI Academy, hosted by Anthony Marshall from IBM’s Institute for Business Value, discusses how business leaders should approach budgeting for Generative AI (GenAI). It highlights that traditional budgeting methods often fall short when applied to GenAI due to its unique characteristics and potential for both efficiency gains and product/service innovation.
The discussion points out that infrastructure, cloud, and data constitute the largest cost components, outweighing the cost of model creation. While investment in GenAI is increasing, funding sources are varied, with only a small fraction coming from net-new spend; most is reallocated from other IT or non-AI budgets. The idea of GenAI being self-funding is explored, but the data suggests it’s not yet fully achievable, although many executives believe it should be.
The episode emphasizes the importance of a long-term view, as GenAI projects currently take longer to yield ROI compared to traditional IT investments. It poses the question of whether to prioritize projects with hard-to-quantify strategic benefits or those with more immediate financial returns. Finally, it identifies three common budgeting mistakes: neglecting bias concerns, underestimating total cost of ownership (including deployment, retraining, and organizational changes), and underestimating integration costs. The video concludes by encouraging a shift in mindset, viewing GenAI as an innovation investment and advocating for a self-funded approach that considers both human and technical resources.
Recommendations
- Address bias concerns from the outset of GenAI projects to avoid major cost blowouts later.
- Take a wide-angle view to understand how Generative AI projects tap human and technical resources to more accurately budget for unexpected costs.
- Rethink what Generative AI investment means to your business today and what it represents to your business tomorrow, to shape how you go about funding and capturing value.
- Consider whether to prioritize projects with hard-to-quantify strategic benefits or those with more immediate financial returns.