Connecting AI Technology with Business Needs
Michelle Bonat is an AI executive known for her transformational work at JPMorgan Chase and AI Squared. She also is an Industry Advisory Board member for Northwestern's MBAi program.
Michelle Bonat is regarded as one of the top 150 global executives transforming the world with artificial intelligence (AI), according to Constellation Research.
She currently is chief AI officer (CAIO) at AI Squared — a software development company — after spending four years leading AI innovation and as AI chief technology officer at JPMorgan Chase. She graduated from Northwestern University with a bachelor's degree and earned her master's degree from the Kellogg School of Management.
Bonat built her career using AI. What interests her now is seeing academic programs that teach how to marry business knowledge with the technology behind AI — insights she developed throughout her career. That's why she was thrilled to be invited to be an Industry Advisory Board member for Northwestern's MBAi Program, a joint-degree program offered between the Kellogg School of Management and the McCormick School of Engineering.
MBAi students develop the skills and experience needed to work effectively with the C-suite as well as machine learning engineers and data scientists.
Bonat recently took time to talk about her career, the state of AI today, and her excitement for the MBAi program.
What excites you about AI today?
I am most excited about the democratization of AI. The way AI has become woven into the fabric of the everyday life of normal citizens, and the pace at which this has happened, is stunning. I’m also thrilled that computer resources are more prevalent and more affordable.
What concerns you about AI today?
As a community, we need to be more transparent and have better education around AI so people are not afraid of it. How does it work? How does it arrive at answers? What is it good at and not good at?
I heard a story about an attorney who used ChatGPT to research a court case. The algorithm made up precedent cases that were then cited in court. This had really bad results and the attorney’s credibility was diminished.
Of course, large language models (LLM) like ChatGPT can hallucinate. There are known steps that can be taken to reduce hallucinations. For example:
- Fine tuning - customizing the behavior of the LLM for specific tasks and contexts
- RAG (retrieval augmentation) - optimizing the output of an LLM by referencing data outside its knowledge base before generating a response
- Advanced prompting - knowing how to query LLMs for more accurate responses
I’m concerned that the take away will be that "AI is bad and makes things up" when there could be more educational awareness of how to use LLMs for best results.
You’ve authored a step-by-step guide to creating an organization's AI playbook. Why is it important for you to share your knowledge and resources in that way?
There are so many choices and ways to use AI for your organization. This AI playbook helps you plan out your AI strategy and in doing so, helps you determine if your AI strategy should be geared toward more of a boat (advancing your position) or a moat (defending your position).
I am often asked how an organization can create their AI strategy and operationalize it. My AI Playbook outlines 15 steps you can take to strategically optimize AI in your organization. Some of the topics the guide covers include business objectives and use cases, collaboration, data strategy, ethics governance and responsibility, intellectual property, knowledge sharing, lifecycle, model selection, optimization, privacy, roadmap, risk, security, and training.
I am a believer that AI is a team sport and we all need to play it together to be most effective. The learnings in our community that are shared in our community is what will make us continue to innovate and to succeed in AI.
What differentiates MBAi graduates when they enter the workforce?
There are three main capability areas MBAi graduates learn in the program that they can leverage in the workforce.
- Prioritization of AI through decision frameworks - This means wading through the hype to focus on the realistic opportunities and challenges. MBAi graduates don't chase the latest shiny object.
- Leveraging business judgment with technical credibility - MBAi graduates can connect business goals and objectives with practical, technical realities and capabilities. They have experience working with technology and not just talking about it. They understand what the technology can offer in a given circumstance and what that means for the business product at hand.
- Solving important problems - MBAi graduates are not looking to plug AI into any hole they can find. These graduates know how to leverage AI to solve certain problems more effectively. They bring their AI knowledge and business expertise together to solve business problems that matter.
What advice would you give a prospective student considering MBAi?
I saw a great quote posted on the board at a conference. It said, “eventually, all product managers will be AI product managers, and all products will be AI products.” Whether you are a product manager, a technologist, or an executive leader, the comment applies. All businesses will become AI businesses, so students should get educated on how to make these future AI businesses successful.