AI is entering a more demanding phase in the enterprise. Enthusiasm and experimentation are no longer enough. Leaders must demonstrate tangible results.
AlphaSense’s data reflects that shift, with mentions of AI ROI in revenue records increasing 116% year over year, and mentions of successful AI implementations increasing 83%. Even major companies are making difficult choices about which features are worth keeping. OpenAI’s recent decision to shut down video generator Sora is one example. Sola demonstrated what technology can do and brought experimental value, but he also highlighted the gap between technological advancement and business practicality.
However, despite rising ROI expectations, many companies still use a narrow definition of value centered around time savings, workflow speed, and task automation. That lens misses where AI creates the most value. In research, strategy, and other high-value knowledge work, AI often improves outcomes by helping people think more clearly, spot gaps, and make better decisions, but not all of this is necessarily demonstrated through speed.
To understand the true enterprise impact of AI, companies need to go beyond efficiency and measure how AI improves the quality of judgment and decision-making.
Traditional ROI metrics are too narrow for today’s AI era
Traditionally, the ROI of enterprise technology has been measured through easily quantified metrics such as time savings, cost reductions, and process acceleration. These metrics remain important, especially in automation-heavy use cases where AI replaces repetitive or stable tasks.
But generative AI creates value in a different way. AI often contributes by improving the quality of results rather than reducing the amount of work or time it takes to complete. AI can challenge assumptions, reveal blind spots, and surface risks that can significantly improve decision-making. You may be able to do all of this without reducing your work time.
These outcomes are harder to quantify than time savings, but more closely tied to real business value. If companies define AI ROI too narrowly, they risk underestimating the use cases where AI has the strongest strategic benefits.
Companies need more sophisticated ways to measure impact
AI is not a one-size-fits-all support. Rather, it is increasingly personalized to an individual’s goals, interests, workflow, and behaviors. Two people can use the same AI system and receive different outputs and still both gain valuable experience.
One example of this change is the rise of customizable AI agents. With AlphaSense, these systems allow users to build customized agents that monitor specific industries, topics, and themes and receive information in a format tailored to how each person works. Approximately 40% of AlphaSense users who created custom agents engaged with scheduled alerts, suggesting meaningful adoption and recurring value. This usage shows that people are willing to invest time in configuring AI to suit their needs if the value is clear.
As AI becomes more proactive, companies need to ask more substantive questions. Did the system surface something the user hadn’t considered yet? Did it improve the direction and quality of the analysis? Did it increase confidence in the decision? Did it influence course correction before the mistake occurred?
In a structured environment, these signals may be easier to track. In areas such as research, strategy, and market intelligence, value can be gradual and less visible. This makes measuring AI ROI difficult, but also very important to define correctly.
As personalized and proactive AI becomes central to daily operations, businesses need an ROI framework that captures the ongoing impact on decision-making alongside individual efficiency gains.
Also read: AiThority Interview with Glenn Jocher, Ultralytics Founder and CEO
The best AI systems expand your horizons
At the same time, as the system becomes more customized, the scope of the system may become narrower.
Optimizing only for relevance based on a user’s past behavior, AI can enhance existing views while filtering out unfamiliar and important information. While this system may feel effective, it actually reinforces blind spots and filters out important information that is unfamiliar.
This is especially dangerous in sectors where powerful decision-making is based on wide-ranging perspectives, signal detection, and exposure to opposing views. The best advisors help people find what they’re looking for, and in turn, help them question whether or not what they’re looking for is right. This level of thinking requires AI measurements to have breadth beyond relevance.
Companies need to start considering metrics such as diversity of information sources, broadening of perspectives, and influence on decision-making. Are users looking at a healthy range of sources and perspectives over time? Has the AI faced risks or counterpoints that were overlooked? Did its results lead to new watchlists, new areas of research, or new strategic questions? Has it substantially sharpened users’ thinking? Are teams adjusting their strategies based on AI input?
This is an important shift for companies to realize the full business value of AI. The most effective AI systems are designed to make decisions that are more informed, more rigorous, and less susceptible to blind spots.
Measure what matters most
Enterprise AI has reached a tipping point where it needs to more clearly prove its value. As a result, companies need to think beyond ROI, which is centered around time savings.
As AI becomes more personalized and proactive, leaders need to measure speed as well as improved judgment and decision-making. This evolution sharpens your thinking, broadens your horizons, and produces better results.
Also read: The infrastructure war behind the AI boom
[To share your insights with us, please write to psen@itechseries.com ]
