What you'll learn:
Organizations are using generative artificial intelligence for “small T” transformations to help evolve their businesses. This means building capacity and managing risk at each stage of the three-stage risk slope:
- Level 1 focuses on low-risk, discrete tasks, such as email management or meeting summaries.
- Level 2 applies AI to specific roles such as coding and customer support.
- Level 3 integrates autonomous AI into products and operations.
Smart organizations under pressure to leverage generative artificial intelligence are realizing that what works in the pilot phase does not necessarily translate into large-scale implementation.
Instead of embarking on a fundamental redesign of key business functions, these organizations are pursuing a series of “small T” changes aimed at increasing value-add, moving them up the risk slope, according to a study by MIT Senior Lecturer Sloan. and Published by MIT Sloan Management Review.
These small-scale transformations are also well-suited to managing the risks of generative AI, such as data security, AI ethics, and compliance challenges.
“Smart leaders are taking a more deliberate and systematic approach to arriving at the big challenges they want to implement using generative AI,” Westerman said in a recent webinar detailing his research. “Every step of the way, they learn about risk management methods and tools and build their capabilities to move toward greater opportunities.”
Climbing the generative AI risk slope
Webster and Westerman defined three categories of AI transformation that represent different levels of risk. Here's how they describe their journey.
Level 1: Personal Productivity
This is the point at which most companies are on the maturity curve. At this level, organizations enable employees to utilize generative AI for low-risk, basic tasks related to their specific roles, while maintaining a human connection during the interaction. One common use case is inbox management, such as using generative AI to summarize emails, draft replies, and flag priorities. Employees also use generative AI to create real-time transcriptions and meeting summaries, optimize daily calendars and automated meeting schedules, and prepare for briefings with easy overviews of markets, articles, and staffing levels. Many desktop tools now integrate extensive language modeling capabilities that improve personal productivity.
In more advanced scenarios at this level, companies use generative AI to recast communications in a different voice or adapt them based on cultural norms. Some leading companies, such as McKinsey, have built company-specific LLMs that give employees access to the company's vast intellectual property resources, allowing them to improve quality and perform tasks more efficiently.
Level 1 work prepares you to do even more with generative AI. “These tasks make people comfortable and alleviate some of the fear,” Webster said. “Then you can move on to level two tasks that start transforming the way your organization operates.”
Level 2: Specialized roles and tasks
At this stage, companies apply generative AI to specific tasks in their roles and business processes, such as coding and data science, human-involved customer support, and low-risk content generation and personalization. Software development is a particularly popular field, and generative AI can give programmers an edge in writing and reviewing code, creating documentation, and performing data analysis.
Generative AI is also reshaping some customer service and sales workflows. For example, researchers say CarMax uses LLM to summarize reviews in hours, rather than having multiple employees working for weeks on end. Other companies are generating personalized scripts for sales calls or using advanced chatbots to handle common customer questions while routing more complex questions to human agents.
“A common theme we see in Level 2 applications is humans and AI working together, finding places where AI can support humans, and humans overseeing the AI’s work,” Webster said.
Level 3: Products and Processes
This is when companies start adding more autonomously generated AI capabilities to their products, customer-facing experiences, and internal operations. Webster and Westerman believe that today's companies are using generative AI as part of a multifaceted toolkit that includes a variety of technologies and talent.
Organizations like Adobe, SAP, and Workday are beginning to use generative AI to take advantage of integrated capabilities to help with rapid content creation, automate marketing campaigns, and provide more sophisticated chatbots that make decisions and perform tasks independently.
Organizations already using AI tools can often start enabling these Level 3 capabilities, Westerman said. “These can require a lot of capacity development and they can also require a lot of risk management,” he said. “That's why companies are taking a cautious approach to reaching this stage.”
Lead an AI-driven organization
Face-to-face at MIT Sloan
Register now
Achieve small changes with generative AI
New technologies improve possibilities at each of the three levels. AI agents are an emerging field of interest, with autonomous task execution to streamline workflows and solve new business problems. This technology spans a continuum from simple AI assistants to autonomous agents that operate on their own based on guidelines and human input. Agenttic AI, a multi-agent version of the technology, would require an “AI manager” to oversee other specialized agents performing individual tasks, the researchers noted.
Although there is plenty of enthusiasm for generative AI and AI agents, there is still a fair amount of skepticism. This is another reason why the “small t” approach makes sense. Webster and Westerman shared the following recommendations for leaders implementing generative AI tools.
- Don’t treat every organizational problem as a nail that can be smashed with a generative AI “hammer.” Focus on the problems this technology can most help solve.
- Think about where your company is on the risk slope and build management buy-in on a plan for moving forward.
- Don't push technology on everyone. Find people who are excited about it and use their enthusiasm and success to drive change.
Bottom line: Take your time. “Building the right strategy is at odds with the gold rush mentality that is happening right now with generative AI,” Webster said. “Take a hard look at your work and business needs and develop your capabilities before moving in a big direction.”
Watch the webinar: Scaling generative AI — getting big value from small efforts
This article is based on a webinar and survey conducted by Melissa Webster and George Westerman, published by MIT Sloan Management Review.
melissa webster He is a senior lecturer in management communication at the MIT Sloan School of Management. She teaches oral, written, and interpersonal communication. Use data to persuade. teamwork. And leadership. She also investigates the adoption and impact of ChatGPT and other generative AI tools in both the professional and educational sectors. Her research investigates the use of generative AI by knowledge workers and its integration into business education.
George Westerman He is a senior lecturer at MIT Sloan. He helps executives understand the transformative potential of AI and other rapidly changing digital technologies. His research-based insights show the questions leaders should ask and the steps they can take to help their organizations thrive. His research on digitally enabled culture and workforce transformation provides critical insights for moving from transformation projects to transformation capabilities.
