IT is a journey
If you’re worried that your organization is lagging behind in its AI initiatives, don’t feel too bad—almost everyone is still in the learning and piloting stages. And if it’s unclear whether there will be a return on investment at the end of all this work, there’s also a financial advantage, research from MIT reveals.
There are at least four logical stages in the advancement of AI, and most enterprises are still working in the experimental and pilot stages, concludes an analysis of 721 companies by the MIT Center for Information Systems Research (CISR). As AI continues, there is now evidence that overall financial performance also improves.
Most of the enterprises in the survey were in the first two stages of AI maturity and had financial performance below the industry average, according to the report’s authors, led by Peter Weill and Stephanie Woerner, both of MIT. Enterprises in the third and fourth stages, on the other hand, had financial performance much higher than the industry average – over 10 percentage points.
Weill and Woerner identified and measured the following four stages of AI progress:
Stage 1: Experiment and prepare (28% of organizations). “At this stage enterprises focus on educating their workforce, formulating AI policies, becoming more evidence-based and experimenting with AI technologies to become more comfortable with automated decision-making,” the researchers explained. Company leaders begin to look at how to address concerns such as ethics and skills to ensure a smooth path forward.
Companies in Stage 1 averaged 9.6 percentage points below industry average, the study found.
Phase 2: Building pilots and skills (34%). In this phase of AI, proponents “define important metrics, begin to simplify and automate business processes, and develop the enterprise capabilities they’ve learned.” In this phase, use cases are piloted, with work on leveraging enterprise data and developing APIs. Work with large language models also begins at this stage.
Companies in Stage 2 averaged 2.2 percentage points below industry average.
Phase 3: Developing AI-driven ways of working (31%). At this stage, AI essentially becomes industrialized, meaning it is available and repeatable across the enterprise. This includes working to build a core platform for AI, providing transparency to decision makers through dashboards, and ultimately transforming the organizational culture to encourage data-driven and innovative thinking. Foundation models and small language models are introduced and applied to entrepreneurial opportunities.
Companies in Stage 3 averaged 8.7 percentage points above industry average.
Phase 4: Get ready for the future of AI (7%). At this stage of achievement, “AI is involved in all decision-making throughout the enterprise,” the researchers say. “They use proprietary AI in-house and many sell new business services based on this capability, AI-as-a-service capability, or both to other enterprises.”
Companies in Stage 4 averaged 10.4 percentage points above industry average.
Successfully moving through these stages of AI growth requires a collaborative cross-enterprise effort, as the technology can reshape and accelerate many parts of the enterprise. Weill and Woerner cite examples of well-known companies at various stages of their AI journeys, such as Kaiser Permanente in the process of identifying AI values and ethics, to DBS Bank committing to a thousand AI experiments per year, which has led to 350 AI Use Cases. And here’s the key – DBS expects the economic impact of these to exceed $1 billion by 2025, they report.
One thing is clear; AI success is a journey, and the ability to quickly use and adapt resources and technology is essential – as new technologies and capabilities continue to emerge almost daily.