Matt Aslett's Analyst Perspectives

The artificial intelligence and machine learning landscape was profoundly altered by the emergence of generative AI into the mainstream consciousness during 2023. The widespread availability of GenAI models and cloud services has lowered the barriers to individuals and enterprises engaging with AI for various use cases, including generating content, querying data, writing code, preparing data for analysis, documenting data pipelines and using software products more effectively. The impact that GenAI has had on enterprise investment in AI is considerable.

While the level of adoption varies across different industries, ISG’s AI Buyer Behavior study indicates that, on average, enterprises expect that almost one-half (49%) of AI spending in 2024 will be associated with GenAI. To put this figure into perspective, it is worth remembering that although ChatGPT emerged in late 2022, only those enterprises at the extreme bleeding edge of AI adoption would have concrete plans to commit any IT spending to GenAI in 2023.

The rise from nearly zero to almost half of AI spending has undoubtedly impacted investment in traditional predictive AI. Although ISG’s AI Buyer Behavior study indicates that 51% of enterprise AI spending will be on predictive AI this year, the study also finds that overall investment in AI has not kept pace with the growth of GenAI adoption. The average proportion of IT spending allocated to AI is expected to grow from 2% in 2023 to 3.7% in 2024. With almost half of that spending now allocated to GenAI, the average proportion allocated to predictive AI has effectively declined.

There is a great deal of justified excitement about the potential for GenAI to enhance many aspects of business processes. As my colleague David Menninger recently explained, however, enterprises should temper this enthusiasm with realism, particularly about the level of investment needed to ensure the skills and expertise required to benefit from the opportunities that GenAI and traditional AI/ML can provide.

As I previously noted, while natural language interfaces powered by GenAI reduce the need for technical and domain expertise to query data, they do not reduce the value of domain expertise for interpreting results. While many enterprises are still validating the potential benefits of AI with proof-of-concept projects, earlier adopters have already learned that deploying AI at scale requires not only AI and ML development tools but investment in infrastructure and software to support AI development and deployment.

Our research indicates that to be successful, enterprises need to be cognizant of the investment requirements for AI-ready infrastructure and AI-enabled data as well as AI model deployment, management, monitoring and integration. Our 2024 AI Platforms Buyers Guide illustrated the need for enterprises to evaluate products that include the ability to prepare, deploy and maintain AI models, including functionality for accessing and preparing data used in the modeling process; exploring, comparing and optimizing models developed using different algorithms and parameters; and governance and monitoring frameworks to ensure that models comply with both internal policies as well as regulatory requirements.

That we also produced the complementary 2024 MLOps Buyers Guide and 2024 GenAI Platforms Buyers Guide illustrates the broad range of enterprise requirements for AI/ML. Our ongoing research highlights the need for investment in expertise and tooling to support the orchestration and governance of AI capabilities across an organization, including financial, safety and security, ethical and ownership considerations in addition to ensuring alignment with business strategies and compliance with regulatory requirements. I assert that through 2026, model governance will remain a significant concern for more than one-half of enterprises, limiting the deployment—and therefore the realized value—of AI and ML models.

To get the most out of investment in AI/ML, enterprises must involve people in business and executive roles outside of the IT department to determine appropriate use cases and success metrics. ISG’s AI Buyer Behavior study indicates the role that people in business and executive roles have in AI technology investments. The results show that while IT, procurement and line-of-business executives are most likely to be involved in evaluating AI product and service offerings, executives such as chief executive officers, chief strategy officers and chief information officers are responsible for determining the overall AI strategy and have a significant influence over final AI technology decisions. This demonstrates the need for software and service providers to address the requirements and concerns of senior executives as well as IT staff and data scientists.

AI involves the development of systems and software capable of automating tasks that have previously required human intelligence. Although much of the current attention is on GenAI, I recommend that enterprise AI strategies encompass ML and deep learning as well as GenAI to deliver capabilities including predictions, recommendations, personalization, speech and visual recognition, translation and summarization. Enterprises should also seek out AI platforms with the full spectrum of capabilities required to support AI development, deployment, management, monitoring, integration and governance. Look for software providers with the expertise and partnerships to address the needs of both IT and business leaders.

Regards,

Matt Aslett