The bottom-line benefits of second chance hiring

Federal spending was responsible for the 2022 spike in inflation

Making generative AI work in the enterprise: New from MIT SMR

Credit: KrulUA / iStock

Ideas Made to Matter

Artificial Intelligence

Generative AI, smart KPIs, and discovery groups: New from MIT Sloan Management Review


With artificial intelligence seemingly ever ascendant, the following new insights from MIT Sloan Management Review explore the most effective ways to use generative AI and the benefits of optimizing key performance indicators with AI. They also discuss setting appropriate sustainability goals and using discovery groups to drive success in addressing diversity, equity, and inclusion.

Make generative AI work in the enterprise

Generative AI is transforming entrepreneurship and improving productivity for workers at all skill and experience levels. While managers can’t be expected to keep track of every technological advance or cautionary tale, they should understand how to make generative AI work in the enterprise.

MIT Sloan senior lecturer led a three-part webinar series exploring this topic in depth.

  • Generative AI Demystified provides an overview of “what gen AI is and isn’t, how it relates to current and past technologies, and what it takes to make good decisions about its use,” according to MIT Sloan Management Review editor-in-chief Abbie Lundberg. The presenters assert that generative AI’s output is highly valuable for summarization and personalization but much less so when a wrong answer can be costly.
  • Finding Transformation Opportunities With Generative AI explores four specific business opportunities that AI enables: a more intelligent customer experience, an improved employee experience, more efficient operations, and the creation of new business models. The discussion also highlights the greatest risks that enterprises face — namely, inaccurate data and privacy concerns.
  • Preparing Your Organization for a Generative Future features Tom Peck, an executive vice president and chief information and digital officer at Sysco. Peck discusses how the food service company uses generative AI to support sales and the supply chain. He also suggests that companies buy AI models instead of building them and argues that CIOs should lead generative AI deployment.

5 trends that will shape AI in 2024

Thomas Davenport at the MIT Initiative on the Digital Economy and co-author Randy Bean outline five AI trends they expect to see throughout 2024 as the technology moves to the forefront of business strategy and executive decision-making:

  1. Though most executives believe that generative AI will be transformational, fewer than 6% have deployed the technology to production because they have yet to see its value.
  2. Companies are rolling out data science models such as machine learning operations systems, which monitor whether models are still making accurate predictions.
  3. Leaders are divided on whether the term “data product” refers to AI, analytics, and data assets in one package or whether it means data assets alone.
  4. The data scientist role is “less sexy” thanks to the emergence of automated machine learning tools as well as ancillary data engineering roles.
  5. The CIO seems to be “in” again as other C-level functions that focused on data, analytics, and AI are increasingly managed by a single technology leader who reports to the CEO.

Read: 5 key trends in AI and data science for 2024

6 shifts to support the transition to smart KPIs

As legacy key performance indicators fall short in delivering the information and insights leaders need, firms are shifting to smart KPIs powered and influenced by AI. About one-third of enterprises use AI to create KPIs today, and 90% of those companies have seen KPIs improve, realizing benefits such as increased collaboration among employees, improved efficiency, and better predictions of future performance.

Michael Schrage, a research fellow at the MIT Initiative on the Digital Economy, and his co-authors note that this won’t come easily. Leaders must juggle the need to improve existing KPIs, create new ones, and understand how KPIs relate to each other. Getting this right requires six mindset shifts within the organization:

  1. Redefine performance rather than assess progress toward predetermined targets.
  2. Model dynamic predictions that adapt to change better than static benchmarks.
  3. Define metrics with the help of AI, as well as with market trends and business objectives. 
  4. Refine KPIs at the enterprise level, not as individual — and isolated — metrics.
  5. Interact with data using generative AI instead of merely reviewing dashboards.
  6. Build strategies for optimizing KPIs and using them to guide decision-making.

The report also examines the algorithms that make KPIs smarter, the governance strategies needed to manage smart KPIs, and lessons learned from enterprises, including Google, Maersk, Schneider Electric, and Wayfair.

Read: The future of strategic measurement — enhancing KPIs with AI

Get to the root of problems with discovery groups (not surveys)

When companies want to hear from employees, they typically conduct surveys. These are rarely effective, writes MIT Sloan lecturer Most employees, skeptical that change will occur, deem it safer to tell executives what they want to hear. Companies get better feedback from discovery groups, Lazu writes in an excerpt from her new book, “From Intention to Impact: A Practical Guide to Diversity, Equity, and Inclusion.”

Anonymous discovery groups go beyond check-the-box surveys to facilitate dialogue among employees, customers, and other stakeholders who are closest to and most affected by the root causes of problems and best positioned to come up with solutions. This ensures that improvement efforts — especially those focused on addressing diversity, equity, and inclusion — are based on what excluded or marginalized groups ask for as opposed to what leaders think they need.

At the same time, those actions should be rooted in larger efforts to improve business performance. Beyond hosting discovery groups, leaders must be willing to reexamine how they recruit and retain talent, commit to cultural change, and develop new strategies for communication and product development.

Read: Gather robust employee feedback with discovery groups

5 steps to set ambitious but realistic sustainability goals

Initiatives like reducing carbon emissions are laudable, but leaders must walk the line between goals that are too weak to be effective and those that are too complex to succeed. This is especially true of goals linked to indirect emissions over which a company has limited control.  

To that end, MIT Sloan senior lecturers  and  provide five steps to set and manage “appropriately ambitious” sustainability goals:

  1. Engage stakeholders to determine the highest priorities. This must balance the desires of external influences (from customers to communities to policymakers) with the perspectives of the employees most likely to lead sustainability initiatives.
  2. Develop an aspirational vision informed by science. Basing a vision on existing, established targets helps focus a company’s efforts on where it is best positioned to make a difference.
  3. Set goals that will gain early traction. These should begin with small, achievable steps that build to larger goals while minimizing the tension associated with change.
  4. Build momentum, and look toward the vision. It helps to seek input from business unit managers who are also early adopters and willing partners. Personal visits from corporate leaders to witness and celebrate successes can go a long way toward reinforcing support for such initiatives.
  5. Scale efforts to lead systemwide change. Leaders must be ready to engage with business partners and take part in policy debates. These conversations should center on what shared success and systemwide change will look like.

Read: Set ambitious but realistic environmental goals

For more info Zach Church Editorial & Digital Media Director (617) 324-0804