Investment in AI/ML technology and talent continues to grow and is proving resilient to economic headwinds and cost cutting, according to the latest Verta Insights research study, which included input from more than 460 AI and machine learning (ML) practitioners.
Some key findings from the 2023 AI/ML Investment Priorities study:
We conducted the study from mid-December 2022 through mid-January 2023, before the current string of layoff announcements from companies in the tech sector and elsewhere. However, even as they have been announcing layoffs, many tech companies have doubled down on their commitment to growing AI/ML.
The recent announcements from Google and Microsoft, for example, spotlighted that both organizations continue to see AI/ML as a growth area. Microsoft’s $10 billion investment in OpenAI demonstrates just how committed the company is to this space.
Media frenzy around generative AI aside, the resilience of AI/ML budgets indicates that business leaders at companies across industries have come to view AI as a competitive differentiator. AI-forward organizations are using AI to grow sales by launching new intelligent products or services, protect revenue by improving customer service, increase efficiency by automating processes, and reduce risks (and costs) by detecting and preventing fraud.
Growing Investments in MLOps and ModelOps
All these use cases are driving investments to support digital transformation and innovation, so the priority given to investments in AI/ML technology and talent make sense. It also makes sense that the research results showed increasing priority for investments in MLOps and ModelOps tools.
The study asked about companies’ investment priorities for 2022 and 2023 across six different categories of spend. Most of the categories were relatively stable from 2022 to 2023, including AI innovation technologies, cloud migration and modernization, data-related tools and infrastructure, and statistical modeling and analytics modernization.
MLOps and ModelOps platforms, however, saw an 8 percentage point uptick between 2022 and 2023, from 35% of respondents citing it as a 2022 priority to 43% for this year. As these are downstream of data and training/experimentation, the results suggest that more mature organizations — that have already made foundational, prerequisite data and infra investments and put in place tools to build and train ML models — are now ramping up investments in technology to operationalize those models.
MLOps and ModelOps tools are essential at this stage and provide capabilities necessary to safely scale up ML. These tools are particularly important as a company expands its use of real-time/low-latency use cases, where a customer interacts directly with the model and expects an immediate response. (Sidenote: Nearly 80% of respondents said they expect real-time/low-latency use cases to increase or increase significantly in the next 3 years.)
Increasing Focus on Operational AI
Ramping up investment in downstream tools and technology can be challenging. Supporting the operational side of AI/ML (Operational AI) is complex, involving multiple stakeholders across multiple functions (Data Science, Data Engineering, ML Engineering, IT, Governance/Legal). The tooling also must support compliance with a growing number of Ethical AI, Responsible AI, security, and regulatory requirements.
In addition, Operational AI requires adopting a mindset that instills operational excellence across the entirety of the ML lifecycle. This requires moving from the bespoke and iterative nature of model R&D to building automation, high reliability, resilience and incident management into the ML lifecycle, supported by the appropriate tooling.
The 2023 AI/ML Investment Priorities study suggests that mature, AI-forward organizations are already focused on this next stage of their digital transformation journey and are investing in the operational side of AI.