Enterprises can achieve significant value from artificial intelligence (AI) through targeted, small-scale initiatives, according to new insights from the MIT Sloan Management Review. The report emphasizes the importance of practical applications of AI across three levels: enhancing individual productivity, integrating AI into specific tasks, and automating production processes. The Vanguard Group, a prominent asset management firm, has reported an estimated ROI from AI of nearly $500 million, citing successful implementations such as improved call center operations, personalized adviser summaries, and a 25% increase in programming productivity. However, only 47% of business professionals believe that their organizations’ AI policies adequately reflect the realities of their work environments, prompting calls for a more decentralized approach to AI governance.
In a recent webinar, MIT Sloan senior lecturers Melissa Webster and George Westerman shared observations about enterprises striving for transformative outcomes with AI. They noted a lack of companies achieving sweeping changes but highlighted that effective leaders extract substantial value from small-scale AI efforts. Such efforts should be systematically deployed at three levels: first, by creating a safe environment for employees to enhance productivity. Common applications include managing emails, transcribing meetings, optimizing calendars, and preparing briefings. Additionally, context-sensitive writing adaptations can aid communication across cultural divides.
The second level involves incorporating generative AI into well-defined tasks. For instance, developers can leverage AI to assist in coding, data analysis, and documentation. Sales and call center representatives benefit from AI agents that provide rapid responses to frequently asked questions, while design teams can generate proposals and visualizations based on minimal input. The third level focuses on automating broader operational processes. AI’s capabilities can support entire marketing campaigns, streamline supply chain management, and identify skills gaps within organizations, all enhanced by conversational AI interfaces.
To derive value from AI strategy, leaders are encouraged to balance immediate actions with long-term vision, effectively “building the scaffolding” for future advancements. Alignment with core business capabilities is essential; without it, pilot projects risk stagnation. This approach underscores the potential of small efforts to yield significant returns.
The Vanguard Group’s substantial ROI from AI highlights its practical applications in streamlining operations. Investments in AI have reportedly improved both efficiency for contact center staff and the services provided to clients. Vanguard’s AI initiatives include agents that empower representatives to access internal information and resolve client issues more swiftly. Additionally, autogenerated summaries keep advisers informed about market perspectives, and AI-assisted code generation has enhanced programming productivity by 25%, cutting the system development life cycle by up to 15%.
Yet, not all AI projects are ready for broad deployment; Vanguard continues to evaluate performance metrics and utilization. The firm takes pride in its training initiatives, with half of its employees completing programs through the Vanguard AI Academy.
Understanding the mechanics of large language models (LLMs) is critical for executives navigating AI’s integration into their enterprises. According to MIT Sloan professor Rama Ramakrishnan, a basic grasp of LLM functionality is essential for informed decision-making. In a recent analysis, he addressed common executive inquiries about AI, clarifying that models can answer questions about events beyond their training cutoffs only if they have access to real-time data. Furthermore, uploading documents within AI prompts does not guarantee that responses will be restricted to that content, as models often draw from a broader training dataset.
Ramakrishnan also cautioned that while modern models can process large context windows, including entire books, excessive or irrelevant information can detract from performance. He noted that hallucinations—erroneous outputs—cannot be entirely eliminated and recommended using secondary models for verification or focusing on structured tasks that are easier to validate.
As organizations adapt to AI, a critical observation has emerged regarding the governance of AI use. Robert C. Pozen, a senior lecturer at MIT Sloan, and Gentreo CEO Renee Fry argue that decision-making should be decentralized. They draw parallels to the early days of the internet, where centralized approval processes stifled innovation. Current practices reveal that only 47% of professionals feel that their AI policies align with operational realities. Such misalignment can lead employees to either seek unapproved channels or disregard the AI tools available to them.
Executives must recognize that decentralization does not equate to relinquishing responsibility. Although leadership should establish overarching policies concerning privacy, security, intellectual property, and ethics, it is vital for team leaders to determine the practical applications of AI within their departments. This localized judgment ensures that AI deployment aligns with the day-to-day activities of employees, fostering a more integrated and effective use of AI technologies.
The evolution of AI governance and its practical applications will continue to shape the future of enterprise operations, as organizations strive to achieve the balance between innovation and oversight.
For more on AI initiatives at Vanguard, visit their official website at Vanguard Group. For insights on AI policy development, explore the resources available at MIT Sloan Management Review.
See also
Utah Launches AI-Powered Prescription Renewals, Enhancing Medication Access and Efficiency
Boston Dynamics Launches Production-Ready Atlas Robot for Industrial Use at CES 2026
Google and Character.AI Settle Landmark Lawsuit Over Teen’s Suicide Linked to Chatbot
China and Singapore: Divergent AI Impacts on Labor Markets Reveal Urgent Trends
Rokid Launches AI Glasses Style: Screenless, Lightweight, and $80 Cheaper than Meta’s Ray-Bans
