What Most Companies Get Wrong About AI Strategy with Jeff McMillan, former Morgan Stanley AI Leader

Jeff McMillan (0:00)

AI does not eliminate jobs, it eliminates tasks. The business model is going to change, the AI is going to disrupt how we do business, and I believe that the companies are going to be successful. They're going to ask the question, how do I rethink my value proposition to the customer?

How do I look at new markets, new products, new services? And how do I expand the pie?

Cindi Howson (0:30)

Welcome to The Data & AI Chief. I'm your host, Cindi Howson. You have a long list of possible AI use cases, but let me guess, you've already blown your budget on unexpected token costs. Meanwhile, your board is pushing you to go faster with AI.

Our guest today navigates this friction every day. Jeff McMillan is the founder of McMillan AI and the former head of firm-wide AI at Morgan Stanley, where he built and deployed AI across the entire company.

Jeff has advised boards and leadership teams on how to implement AI at scale with proven value. Along the way, he has found the same pattern holding most organizations back. In this episode, you'll learn what's actually stalling AI transformation at most companies, and how to start fixing your data foundation, and what AI's economic impact means for leaders. Jeff, welcome to The Data & AI Chief.

Jeff McMillan (1:34)

I'm happy to be here. Thank you.

Cindi Howson (1:36)

We're happy to have you here, Jeff, and I know you travel the world visiting with customers. Where are you joining us from today?

Jeff McMillan (1:45)

I am actually home in lower Manhattan, New York. So no travel today.

Cindi Howson (1:51)

Okay. Well, New York, my former backyard, I miss the big apple.

As you think about your journey to this role, tell us a little bit about McMillanAI and why you decided to leave a well-known brand like Morgan Stanley.

Jeff McMillan (2:11)

So, Cindi, I was with Morgan Stanley since 2009 I actually somewhat accidentally ran the integration of the platforms between Morgan Stanley, Smith Barney, and I became their first Chief Analytics Officer in 2016

Then in 22, we had a small team on the West Coast, and at the time, we met this unknown company called OpenAI. Morgan Stanley became the first customer of OpenAI in financial services. We were the first ones to deploy anything into production not long after GPT. The first version of GPT, ChatGPT, went live.

In 24, I moved into a firm-wide role responsible for essentially effective and efficient deployment of the technology.

And I really believe this, like these firm-wide jobs, they have a shelf life to them. And I really think the way I refer to my job and when I work with companies say is, the chief AI officer is not there to build AI for the company, they are to teach the company how to build AI. They are to build the roles and responsibilities and the infrastructure. And innovation does not happen at the center. And candidly, I got to the point in my role where that work was largely done. The businesses were enabled with the right tools, 80,000 employees. We deployed hundreds of use cases. And the firm was extraordinarily well positioned. And in 2026, there's not a lot of expertise in this space. And I saw a small window for me to really maybe have an impact on how firms deploy this. So I left in January of this year and I've been at it ever since.

Cindi Howson (3:53)

What a great story and journey.

And there's a few things that I'm noting in your story. 2016 is an early time for the title of Chief Analytics Officer, as opposed to at that time Chief Data Officer. So it sounds like you were always focused on the value side of the data equation. Is that a fair assessment?

Jeff McMillan (4:18)

Yeah, I mean, honestly, my first job back in 1996 at Morgan Stanley, I was working on a data project to clean up the customer information right after the long-term capital blow up. And firms were trying to figure out what their exposure was to long-term capital management, LTCM period. And I've had a variety of roles that surrounded data and analytics. But in some ways, my story is the story of the industry. We went from pure data to how do you monetize with basic reporting tools?

How do you then use simple algorithms? And then how to use more sophisticated algorithms? And in the wealth space, it was all about how do you identify the best ideas for each client? We deployed a platform called Next Best Action in 2018, which was the first algorithmically driven idea generation tool for financial advisors. And it really worked, right? When you give smart people better information that's of high quality, that's targeted to the personal needs of their clients, they do a better job than people don't have that information. And it's always been, I think there's too much argument about which is better, the machines or the tech, but at least up until now, everything I've done is shown that, high quality, high performing people with high quality algorithms just produce better outcomes.