2024 marks itself as a pivotal year for Artificial Intelligence in its entirety. Emerging from the groundbreaking developments and specializations of AI in commercial use, industries have witnessed the technology being implemented with growing breadth—signalling the shift from Experimentation to Wide Scale Deployment.

‍

This year’s shift, in turn, has caused a redirection in firms’ AI strategies; now that resources have successfully built and/or employed monumental systems into their AI Portfolios, attention is turning into several pivotal questions:

‍

• How can I make my AI system perform better?
• How safe is it when in operation?
• How can I now manage these risks and the potential drift of the model?

‍

With growing enterprise attention towards AI Risk Management tools, this year has provided Calvin with invaluable insights into the deployment field from our growing clientbase. In this blogpost, we look to share AI Risk Management industry reflections, expectations for 2025, and our perspective on mitigating emerging risks while maximizing AI-driven value.

‍

At a Glance: AI Deployment in 2024

‍

AI deployment has taken center stage in 2024. In a study conducted by IDC and sponsored by Microsoft, key trends such as productivity, advanced solutions, generative AI adoption, and increased returns have emerged as focal points for firms over the past year.

‍

At the forefront, AI projects have demonstrated a measurable impact on business outcomes; for every $1 invested in generative AI, the return on investment (ROI) was impressively reported at $3.7x, with top leaders in generative AI achieving even higher returns at an average ROI of $10.3x. To support this, according to Deloitte's latest research on generative AI in the business landscape, a majority of senior executives surveyed—specifically 58% of C-suite and board members—reported realizing tangible benefits from generative AI implementations. These benefits primarily center upon driving organizational innovation, enhancing product and service offerings, and strengthening customer engagement strategies.

‍

Such aligns with the IDC’s findings that productivity remains the top business goal for AI systems, as 92% of AI-enabled firms were found to leverage AI for enhancing employee productivity. Among them, 43% reported that productivity-focused use cases delivered the greatest ROI. The use cases of customer engagement, topline growth, cost management, and product/service innovation closely follow, having found significant success this calendar year with nearly half of the companies surveyed expecting AI to experience a profound impact in these areas over the next 24 months.

‍

With respect to Generative AI, adoption has surged from 55% in 2023 to 75% in 2024, signaling a shift towards broader adoption and integration. Notably, the highest ROI from generative AI was observed in the financial services sector, followed by media & telecommunications, mobility, retail & consumer packaged goods, energy, manufacturing, healthcare, and education. Our strategic focus on the top three industries has proven highly effective, as we’ve already refined our platform to meet the out-of-the-box needs of these sectors. Effectively, this growth in generative AI will continue to extend, as the next 24 months will likely see companies building custom AI solutions tailored to increasingly niche industry requirements—such as copilots and AI agents.

‍

This year, we’ve witnessed a significant shift toward Retrieval-Augmented Generation (RAG) systems, particularly in the banking and insurance industries. Earlier this year, we introduced our groundbreaking quantitative LLM testing suite, designed to fine-tune these models and deliver unprecedented precision testing to ensure performance, robustness, and bias-free outputs, amongst others. Such reflects a growing sophistication in AI fluency, as companies embrace more complex, tailored solutions.

‍

With these exceptional developments, we anticipate continuous growth in AI adoption as a whole. The IDC highlights that, beyond enhancing business value, 29% of AI leaders implement AI solutions within three months, compared to only 6% of lagging firms. Tools like Calvin are instrumental in accelerating this pace, streamlining deployment, and empowering companies to achieve their AI goals efficiently.

‍

2024’s Array of Opportunities and Risks

‍

While AI deployment has expanded significantly in 2024, technological advancements are accompanied by substantial organizational challenges. In the market, this dynamic has manifested through prolonged implementation timelines, rollback of AI systems, and strategic recalibrations due to model performance inconsistencies and unexpected outputs due to hallucinations.

‍

A Harvard analysis of S&P 500 companies revealed that over 60% recognize material AI-related risks, with organizations identifying comprehensive risk categories including cybersecurity, innovation, regulatory, intellectual property, ethical, and potential reputational risks. Notably, 20% of these companies disclosed exposure to three or more distinct AI-related risk domains.

‍

The research uncovered that more than 30% of companies are concerned, in their strategic assessments, that insufficient AI technology integration could significantly undermine their competitive positioning and lead to lost market share, if unable to offer market-acceptable products and services with AI.

‍

Correspondingly, approximately 30% of organizations expressed concerns about navigating the intricate landscape of emerging AI regulations; over 15% specifically highlighted challenges related to data protection and privacy frameworks within the evolving governance environment.

‍

Lastly, companies are increasingly disclosing the risks associated with failing to use or deploy responsible AI. Drawing from IBM's 2024 research, approximately a quarter of organizations recognized potential reputational risks associated with AI implementation. Around 15% explicitly addressed ethical considerations, while 1 in 5 companies candidly disclosed potential vulnerabilities of flawed outputs, potential bias, or defective/social harming risks in their AI models, algorithms, and training methodologies—signaling a growing awareness of the critical need for comprehensive AI risk management strategies.

‍

Predicting the State of AI in 2025

‍

With AI Risk Management emerging at the forefront of both business and technical line strategies, proven tooling becomes increasingly critical as we approach 2025. This focus is particularly pronounced in the financial services, telecommunications, and mobility sectors, where high-stakes AI deployments intersect with evolving regulatory landscapes like the EU AI Act. As AI technologies transition into a maturation phase, organizations are driving towards optimizing systems that align with competitive benchmarks and industry standards, simultaneously ensuring comprehensive risk monitoring and mitigation strategies that address AI's inherently complex, opaque operational nature.

‍

In a prospective for 2025, Sarah Bird, Microsoft’s chief product officer of Responsible AI, noted, “Even as models get safer, we need to bring testing and measurement up to the worst of the worst threats that we see—testing that represents a sophisticated adversarial user and what they’re able to do”.

‍

At Calvin Risk, we’re committed to offering the best-in-market quantitative solution to tackle AI performance, regulatory, and ethical risk management excellence—making Responsible AI a verifiable reality.

‍

About Calvin

‍

The Calvin software solution is a comprehensive, assessment-first governance platform designed to maximize business value from both predictive machine learning and generative AI applications. Calvin's modular architecture allows organizations to leverage specific functionalities based on their unique needs, effectively serving as an all-in-one suite that can be tailored to varying compliance and risk management demands. The solution is frequently deployed within clients' own cloud environments (e.g. private or virtual private cloud) to ensure data privacy and alignment with internal IT policies. With its unique focus on combining quantitative risk evaluation with governance best practices, Calvin empowers enterprises to understand and manage their AI risk profiles comprehensively, providing enhanced decision-making support for executives.