Ep 61: Reframing Knowledge Work through AI with Houlihan Lokey CIO Allen Fazio

Evan Reiser: Hi there, and welcome to Enterprise AI Innovators, a show where top technology executives share how AI is transforming the enterprise. In each episode, guests uncover the real-world applications of AI—from improving products and optimizing operations to redefining the customer experience. I’m Evan Reiser, the founder and CEO of Abnormal AI.

Saam Motamedi: And I’m Saam Motamedi, a general partner at Greylock Partners.

Evan: Today on the show, we’re with Allen Fazio, Chief Information Officer at Houlihan Lokey, the world’s number one investment bank for mid-cap M&A. Before joining Houlihan, Allen spent 20 years at Disney running IT for Disney Cruise Line, Disneyland, and Disney Imagineering—so he’s seen a few technology revolutions in his time.

There were three things from our conversation that really stuck with me.

First, Allen talked about the orchestration layer problem: what do you put at the center of your AI strategy when there are suddenly 25 different tools competing for attention? Houlihan is betting on training one model on their treasure trove of 20 years of M&A deals rather than fragmented, cross-platform approaches.

Second, the speed of this thing is wild. Four months ago, Allen’s team tracked 60 AI products in use across the firm. Yesterday: 190. That’s not even counting AI quietly embedding itself into every existing tool they already run.

And third—and this one may be surprising—Allen said his summer interns were outpacing full-time associates in AI adoption within weeks. The younger you are, the more you’re using it. Which raises some fascinating questions about how careers in professional services are going to evolve when entry-level work gets automated.

Thanks again for joining today. Maybe to kick us off, do you mind sharing a little bit about your journey in technology—and a little bit about your role today as a CIO?

Allen Fazio: First, thanks for having me. I’m super excited to be here today.

I started my career actually in accounting. I was with a CPA firm, became a CPA early in my career, and everyone’s like, “Well, how’d you end up in technology?” My degree was split half and half—half accounting and information systems, and half MIS and computers.

Then the journey went from public accounting. I spent 20 years with Disney theme parks. I ran IT for the Disney Cruise Line, then moved out to California, became a CIO for Disneyland and Disney Imagineering, and then magically ended up in investment banking.

So for the last eight years, I’ve been with Houlihan Lokey as their CIO, responsible for global IT. When I joined the company was about $700 million; today we’re $2.5 billion. So we’ve been in absolute growth mode for the last six, seven years. And again, for someone who switches industries fairly regularly, I love being in the financial services space.

But along the journey it’s been laundry facilities, help desks, logistics, point of sale—everything you can think of in the Disney world. So it’s been a great journey. And it’s times like today where you pull on everything you’ve ever done to try to figure out what to do next.

Evan: I appreciate you sharing. Maybe for some of our audience that are less familiar, do you mind sharing a little bit about Houlihan Lokey—explain what you do? I think it’s probably a bigger firm, with a bigger impact, than most people realize. It’s probably less known than Disney.

Allen: I had to look them up and go, “What do they do?” So I fully understand when someone—when you say Houlihan Lokey—people say, “Is that a law firm?”

No, we’re a global investment bank. But even that could have several definitions. We specialize in three things.

First, we are number one in the world in mid-cap M&A. So we will do more M&A transactions between $100 million and $2 billion than anybody else on the globe—mergers and acquisitions.

We’re also number one in the world in debt restructuring, period, regardless of size and scale. So if there’s a distressed company or distressed assets in the world, Houlihan is often called to advise either the creditors or the business itself going through a distress situation. Things like Toys “R” Us or Sears—all the way to China and Evergrande. Anyone who’s in distress, we’re one of the first calls usually.

And then third is our valuation group—financial valuation. We value assets, illiquid assets: what are they valued? We work with boards of directors on advisory and assurance.

So our firm is really multifunctional, but those are the three major divisions. Probably on top of that is capital solutions—if you need capital for a transaction, we have a whole group of bankers that are world-class at obtaining the necessary capital to help the transaction through. That’s what we do, and we do it at a very high level for the mid-cap and/or global spaces we play in.

Saam: You’ve been through technology cycles before. Where are we? Help our listeners context-set in this cycle and hype cycle, and what’s your perspective on AI specifically as it relates to investment banking?

Allen: This move to AI has all the same characteristics. You’ve got the veterans in your organization going, “I’ve been doing it the same way forever,” and the young people coming in going, “Wait—this is going to transform the way I do work.”

So we’re right back into that moment of a revolution on our doorstep. However, this one’s going to happen at a speed we’ve never seen before. It took time to buy everyone computers, network them, decide how to do security, and learn not to put your floppy disk next to a magnet. This one’s going to happen at speed—and you’re already seeing that.

Here we are with AI, and it’s coming into the business faster than anyone could ever think. It’s not a thing—it’s dozens of things. We keep track, using our network software, of every AI product used. Four months ago it was 60. Yesterday it was 190.

Now, a lot of them are one-time use—people kicking the tires. But that’s just products that say “AI,” and then AI is seeping into every product we run, every tool we use. It’s not “AI—go buy AI” or “go get an LLM.” It’s coming into the organization not even in trickles—it’s in absolute floodplains right now.

For investment banking: at our heart, we’re a professional services company. We’re like a law firm. We’re like an accounting firm. We’re like a consulting firm. We’re not necessarily like a bank with a trading desk doing stocks and so forth. We act a lot like a professional services company, which means our main asset—the main thing we do—is have smart people do work.

Now, what do those smart people usually do? Excel, Word, PowerPoint, and a lot of analysis. And what is AI really good at? Analysis—bringing together data, etc. So you’re seeing this toolset coming into play that’s going to fundamentally change how the work is done.

Evan: Are there any tips or tricks you’d recommend to help people explain what’s possible—educate what’s possible? Because I think not everyone is on the same page about the real opportunity AI presents. Today we still see a lot of toys—they’re cool—but if you looked at what OpenAI or Anthropic is doing and wrote it off as a chatbot, you might be missing the bigger picture. How do you get people to open their eyes and realize what’s possible—realize the biggest opportunity is not today, but what’s coming on top of the platforms and technologies we’re currently getting exposed to?

Allen: As you’re trying to get your organization to move into the AI space, it’s one thing to go, “Here’s the new shovel—let me replace your shovel with this new powerful shovel.” It’s another one just to get them engaged.

How do you get everyone in your organization to find ways to change little things in their world every day?

This one’s going to sound weird. Our favorite one right now is: we use an AI tool that takes a look at our email, our meetings, and so forth. And our favorite thing right now—one of my VPs came up with this—is every Friday we say, “Could you please summarize my week in the form of a Comedy Central roast?”

They’re brilliant. They’re amazing. They’re funny. They’ll say, “You were in a meeting with Allen—he was pontificating.” I’m like, “I wasn’t pontificating,” but they’re wonderful.

And we share them—we share our summary of the week with our peers. And that little moment leads to: okay, I use the same toolset to say, “Summarize my week, give me all my action items that I agreed to, and all my direct reports and any action item they signed up for,” instead of my normal processes where I forgot about half the assignments I handed out and I don’t follow up on them.

So I am much better on Friday afternoon and Monday mornings than I have ever been before, without any earth-shattering change to the way I do things.

And here’s an example: if you can create these moments where everyone is jumping on the tool for something they find enjoyment in, they like, they see value in—that leads to the next learning, and the next learning, and the next learning.

Now you’re building out an innovation culture, not deploying a technology or a toolset.

Saam: What are two or three use cases for applied AI that you’re particularly bullish about—where your peers at other firms may be underestimating it or not as much believers—but you’ve seen proof points around and think can really be transformative?

Allen: We’ve been looking for almost three years for what I’ll call the orchestration layer. What do we put in the middle of the picture? If you were going to draw a picture of AI and there’s going to be lots of different logos and a lot of different toolsets on that map—what do you put in the middle? What do you put at the core of your AI strategy?

We went out and talked with Oracle. We flew their AI guys down to one of our guys’ houses in Texas for a couple days. We’ve talked with Microsoft on Fabric, and we’ve talked with Google—well, it was AgentSpace; now it’s Gemini Enterprise.

So what do you put in the center? Is it ChatGPT? Is it Gemini? What do you put in the middle? Because you’re not going to want to train everyone on 25 tools.

More importantly, at least for us, whatever we put in the center—we only want one thing to have: we only want to train one model on our unstructured data. We are sitting on a treasure trove of more M&A deals than anybody on the planet. We have a 20-year history of M&A deals and bid sets and industry analysis that no one else has. We’re sitting on this repository. We only want to train one model on our unstructured data.

Number two: if we’re going to build custom agents—and as we go down agentic AI—we don’t want to do that in five different tools. So let’s build our agents in a place, at least to start.

So what do you put in the center so that your unstructured data and your agents have a place to live? And then as soon as you’ve got that squared away, now it’s: what about everything else?

There are tools that are specific to our industry that are emerging. There are generic tools like Grammarly or Freepik that add value, but they’re not core—they’re incidental to some of the things we do.

And as you take your core and you open out, now you go, “Oh, what about third-party data?” Especially in investment banking: here comes Capital IQ and FactSet and Bloomberg and all this rich data. Do we attach to it and bring it into our core, or are there third parties that now have that as part of their knowledge base in their AI toolset? So do we buy it and add it, or do we purchase it and add it on? As we think of AI, it’s all of the above.

Plus, here comes toolsets that are very specific—like for our legal department. That’s a buy. That’s a tool for legal that can add value to NDAs, MSAs—immediately raise the bar—so that some of the low-hanging tasks in legal are addressed by AI, helped along by AI.

The picture gets big quickly with AI. And the real question comes back to: how fast can we bring the business along? How quickly can they move and understand the changes in toolsets we’re bringing?

Because for every toolset we roll out there will be a moment of: “I like X better than Y.” “I don’t trust it.” “I got a bad answer so I’m never doing it again.” Back to the days of Excel—“Oh, I didn’t like it.”

So the innovation mentality—the culture of innovation that has to happen as we bring all these tools to bear, stack them, and the sequence we bring them to bear—is as important as the toolset.

Saam: I think professional services broadly—and investment banking as a subset, but also legal and others—is an area where we’ve seen a lot of early uptake of AI tools. Relative to other markets, we’ve seen much faster uptake in professional services. There’s conversation in the zeitgeist of: what is the future of professional services? One extreme is: AI does 80% of the work and it changes the dynamics dramatically. The other extreme—which I don’t think any of us believe—is it’s a 10% productivity boost but nothing more. Where do you think it lands? What does professional services look like in five years? What does an AI-native investment bank look like?

Allen: It’s a great question, and I don’t think any of us have the answer. But here’s how we think about it today.

We believe a banker with AI is better than a banker. We don’t believe AI is better than a banker with AI. That’s the foundation of where we think we’re going.

Now, that does not mean massive change is not coming.

Let’s use it as an example as a service provider—and this applies to banking, law firms, etc. As we start using AI on my service desk: my entry-level service desk person has access to a knowledge base that is 20 or 100 times better than the one we created five years ago. They have access to everything that’s happening out in the world around a potential problem. That makes my entry-level service desk person smarter and better than ever before.

What’s the ramification of that? Do I need fewer low-entry-level analysts, or do I need fewer high-level analysts because the lower level can do more? Do you stay at entry level longer because of the capabilities you need to move up? Now the bar has raised—what you have to have to move up—because the AI is making you that much better.

I currently have three tiers of help desk. Does that move to two? Does that move to one?

If you talk to people leveraging AI to develop software, you still need that ringer who can bring it all together—but you need one of them, too. Do I become more of a business-analyst-heavy organization versus a developer organization? Will I grow in some aspects and not in others, or shrink in others?

And then you get to: how am I going to train a BA if they don’t have the lower-level tasks to learn?

Same thing in investment banking: how are you going to train and create the next generation of MDs if all the lessons that make you a great MD are in being an analyst and associate and doing this type of work that maybe you don’t have to do anymore? What’s the lesson you have to pull from the work for the knowledge to succeed in the future versus actually doing it repetitively?

I think professional services are all over this because there are mundane tasks that are very time-intensive. If you’ve never been an analyst where an MD drops off a 900-page contract on your desk on Friday night and says, “Can you pull these 30 data elements out of it by Monday?”—AI can help do that. And I’m not sure it hurts anybody if I can solve that problem in minutes versus a weekend.

Evan: I’m curious—if you look back, are there other wins you’ve had that have been unexpected? Or if there’s a CIO listening trying to figure out, “How do I put some points on the board and open people’s minds about what’s possible with AI?”—are there three pro tips or easy wins you recommend to get the ball rolling?

Allen: We’re probably six months away from starting to stack the wins of: “Hey, here are the agents we built. Here’s the workflows that have gotten better. Here’s how we’ve started to install this culture of innovation.” So I wouldn’t say we’ve had a ton of wins—but it’s okay.

The story or analogy I tell is: okay guys, let’s jump in my wayback machine and go back to the late 1800s. We’re all sitting in St. Louis and there are a million covered wagons in town, and every day you’re seeing 50 of them head west, and you’re like, “We gotta get going. We gotta go.”

And then you’re like, “Whoa, whoa, whoa—half of them don’t make it. The other half get there and don’t survive.”

So we’re sitting in town—you can feel the pressure of “it’s time to head west.” We’ve spent enough time interviewing other people, seeing what they’ve done, seeing where the successes are. Now it’s time for us. We’re ready to go. We are stocked up, loaded, ready to go. And we think we have a higher chance of success leaving now than maybe the people who left early, and/or the people who are going to leave after us.

No one will know if we’re right or wrong for a while, but it’s definitely time for us to head west. And we’ve taken the last year and a half to make sure we have enough of what we think are the right tools and provisions to make the journey.

Evan: Are there maybe one or two areas where you feel—looking forward in the short term—where you want to focus?

Allen: Our focus is this: when I get a report, I see what everyone’s trying to do with AI. And believe it or not, the younger you are, the more you’re doing. Shocking.

Over the summer, our analysts were the most users of AI, followed by our associates, followed by our VPs, followed by our managing directors. So literally, the longer you’ve been in your career, the less you’re spending time with this emerging technology—no surprise.

What was great was: in May, when the interns came in, the interns started and I watched their line grow—and then they outgrew the associates, and they started to threaten the analysts. And I was like, “Okay, the interns are all over this technology.”

So for us, our focus is going to be on that level. Let’s start with the analysts, associates, and interns, and let’s really focus on their capabilities.

Where I think some of our peers are—I'm not saying they’re wrong—but where they’re spending more time than I’m going to spend is on these high-end use cases: “Here’s what an MD is doing,” “Let’s see if we can automate an entire industry analysis.” I think it’s too complex—too 400-level class. Let’s go do the 100-level work, then move up to the 200, and be smart about how we’re approaching it.

We’re not going to be able to contain it. The minute you start on the 100-level class, you’ve got the banker coming out of the Big Four accounting firms going, “I want to be able to do this,” and the 400-level action items come at you really quick.

It’s being able to control those—only authorizing the right ones, and a limited number of them—while you work on the 100-level class with the analysts and associates. Because if we can really change the way they’re doing their work, we can then move up to the next level, the next level.

Our fear is: any senior executive, when all of a sudden it doesn’t work the way they think it should, they’ll back away from it—way too far away. “This doesn’t work. This is crazy. I don’t need computers.” You want to avoid poisoning the well by the seniors while the juniors get to learn new capabilities that the senior-level people never had.

You’ve got to make room for that. Back to a culture of innovation: you’ve got to go to the MDs and directors and talk to them about what time and what they have to do to help the analysts and associates, because they have to find room to play—or try and fail. We’re usually on very tight timelines.

If you don’t create the right culture, you’re going to find your analysts and associates doing this on their own—buying their own AI tools, potentially taking client data and running it through something that’s not sanctioned. And that’s where you get into the heavy rules and regulations. The regulation in our industry makes it very fascinating to navigate.

Saam: My impression—80% right and 20% wrong—is: in investment banking, if I join as a new analyst, the way I differentiate myself historically is intense work ethic and almost zero-defect craftsmanship—perfect models, immaculate decks, the 2 a.m. revisions. That’s how I stand out and progress my career.

Now, if I fast-forward and you’ve deployed the 100- and 200-level classes of AI agents and they can do these things for me, one perspective is: how does young talent differentiate and progress if the thing they differentiated on, AI is better at?

What would your response be? If I’m 20 years old and two to three years out from joining a firm like yours as an analyst—and let’s assume AI has progressed and you’ve automated even more of the 100- and 200-series work—what would your advice be? What would you look for differently in hiring me than you would have 10 years ago? And what would you tell me to focus on to advance my career?

Allen: Great question. I would say I’m living proof, going back to the fact that we’re here again.

I came out of college with this accounting-computer thing that no one knew what to do with. You’re going to come out of college, young 20-year-old, and if you do it right, with skills that no one prior to you had.

If you’re going to come into investment banking, the days of being a pivot table expert are coming to an end. You still need to know that. You still need to be good at it. But the new currency for you—that’s going to separate you and be an early accelerator for your career—is how to bring this new technology to bear.

What I mean by that is: if you’re going to do an analysis with AI, we also know you can get a hallucination—maybe it’s not 100% right. So are you going to be able to do the analysis and then check it in ways that are better than the person before you?

It won’t be, “I did the spreadsheet better.” It’s, “I had AI do the analysis. I was able to check it, and I was able to bring back a 99% answer versus my peer who brought back a 96% answer.”

You’re going to see the questions change immediately in the interview process. I thought it was going to be things like R and certain software languages. It’s not—AI has passed all that up.

The mastering of these toolsets—this is an exciting time. You have a chance to not just come in and brute-force your way into an industry.

Wouldn’t it be great if being an analyst in investment banking didn’t mean 90-hour, 100-hour weeks—that you could actually do it in 50, 60 hours and have the same lessons?

So I’m excited if you’re 20 years old. But if you’re 20 years old trying to use the existing playbook that your teachers are probably telling you, you better open up a new playbook. You better bring these new skills to bear, because it’s going to be a race.

There are two components to investment banking.

One, we call the rule of 99. If it’s 99% accurate, is that good enough for an investment banker? If you could be 99% right but 50% faster, is that a trade-off you would do in a pitch? If you’re pitching work, would you make the trade-off of 50% less costly for 1% less accurate? Or would you feel that’s not an equitable trade-off in this industry? I don’t know the answer yet—that will play out.

Number two is regulations, because the regulators are behind in this space.

Think about investment banking—a very simple concept: can we record the meeting? Up till now, if you and I are meeting, we don’t have to keep a record of the conversation. But the minute it’s digitized, it becomes something that must be retained. So all of a sudden now conversations that bankers are in are being recorded, which means they have to be retained—but the rules aren’t defined yet.

Do I have to keep the verbatim of the call? Do I have to keep the recording? Can I keep a summary of the call? But the summary of the call is based on what I asked AI for—and how I ask for the notes can drive what’s in them.

From a regulatory perspective, we’re in a little bit of uncharted territory.

So again, new person coming in—this is going to be interesting: how you navigate the rules and regulations of everything we have. You both know this: every time you go to do a presentation about your company, you have the safe harbor. You have to do the safe harbor rules, right?

If I’m a banker and you say you want to record the call, should I have a 3x5 card where I read a certain statement that says, “Hi, I understand this is being recorded. I’m Allen Fazio from Houlihan Lokey, and I accept the use of this in the following ways.” Am I supposed to do that?

We’re talking to law firms: how does that work? It hurts your head after a while.

So 20-year-old: I am so excited for you. This is a chance for you to absolutely accelerate into a new world where it won’t be the MD stepping into it—most of them. It won’t be the director stepping into it. Some of the VPs will. But you’re now going to be up against a group of associates and analysts that—if you come in with a better skill set—everyone else is going to be over your shoulder going, “What are you doing? How are you doing it?” It’s exciting.

Evan: In the last couple minutes, we love to switch to a lightning round because we’ve got a little time left. We’ve got four or five questions. We’re looking for the one-tweet response to questions that are very difficult to answer in one tweet—so please forgive us in advance. Saam, you want to kick it off?

Saam: Maybe to start: how do you think companies should measure the success of a CIO in this AI era?

Allen: Are they facilitating the firm moving into the new world?

Evan: Allen, one thing I’ve been very impressed by is you’re very up to date on the technology—you’ve thought through it deeply. How do you recommend other CIOs stay up to date on the latest AI trends—technology, applications? Any advice for an incoming CIO trying to get up to speed?

Allen: Probably three things, real quick.

Network with others—leverage everyone. I spend a lot of time with venture capital. I spend a lot of time at dinners with companies and do a lot of pilots. You’ve got to spend the time. Most people are not doing that.

I’m out to dinner three, four times a week. And I would tell the young person: 60% of the time it was a bad night, but 40% you walk home with knowledge—so you’ve got to keep swinging.

Saam: Switching gears to the personal side: is there a book you’ve read recently that’s had a big impact on you, and why?

Allen: A whole bunch. I’m trying to find it here. Ashwin—one of my friends down here, a CIO—just wrote a book on AI governance and how to govern this space.

I’m usually caught between three different books: a leadership book, a technology book, and a science fiction book. I’m usually reading all three simultaneously.

And I was up at Google and I’m reading Snow Crash from 1993. Everyone tells me it’s a classic, but I’m stepping back in time on this one. It’s surprisingly relevant today.

Evan: The fact it’s so relevant speaks to the imagination of Neal Stephenson.

Okay—this doesn’t have to be AI, but it may be: what’s a new technology you’re personally excited about? Doesn’t have to be work-related.

Allen: I’m totally into what I call “leave a legacy.”

My dad’s getting older—we’ve got older parents and things like that. My mom passed a couple years ago. So all the technologies that let my dad leave a legacy—things like Remento—where I can let him tell a story, and with AI under the covers, the ability for him to answer a lot of questions, and have an AI allow my granddaughter to ask him a question when he’s long gone and him answer it with the knowledge—that, to me, is fascinating right now.

Because the clock’s ticking. A lot of us are at that age where parents are in their 80s right now. Now is the time—you can’t wait for new technology.

It used to be heritage, where you’re from, your family tree. This is a whole other level. I call it letting your parents leave a legacy versus just opening up a book and saying, “Here’s a picture of your grandparents you never had a chance to meet.” Those are the technologies I’m most interested in right now.

Saam: What do you think will be true about AI’s impact—generally speaking—ten years from now, that most people would still consider science fiction today?

Allen: I think the biggest impact in ten years will be 30-year-olds or 40-year-olds turning to fresh faces and going, “You know how we used to have to do that, right? You know how we used to have to do it the hard way? We used to have to walk uphill both directions to school—you have it so much easier than us.”

So I think that’ll be in play. But at the same time, it’s never that simple. It’ll be the pain that goes with it: “Oh, you don’t get it back. I did this with AI and it cost the firm $50,000 because we didn’t know what it was going to cost to run it. You guys have access to bigger data sets and more information—we had limits.”

It’ll be the old man’s story to the young face, but done with all the capabilities that came in. I feel like I’ve lived it. It’s exactly what we told people before computers: “Oh, you’ve got computers to do that?” Same story, different timing.

Evan: Final question—I know we’ve got to run. Allen, what advice would you give to up-and-coming technology leaders?

Allen: Run and hide. Marketing is much easier. No—that’s not the right answer.

What I would say is: you have to be at the crossroads of a couple things.

Number one: you have to talk business. You cannot purely talk tech in today’s world. You’ve got to speak in terms the business understands. My advantage was I started as an accountant—I can talk financials. I can convert tech to business.

I tell stories. I deliver almost everything in story format. Even if you’re the most technical and you have the right answers—I work with brilliant people who are much smarter than me—but they cannot convey the message the right way.

So things like storytelling, things like how to communicate—those are critical to your success. People will tell you you have to be technically great, and you do—but it won’t be enough.

Make sure you’re focused on your soft skills. Make sure you’re working on these other components, because at some point there are enough really good technical people. There are too few people who can convert that into a way that the business can consume it, fund it, support it, and keep you in your job long enough to get it done.

Evan: Appreciate you sharing, Allen. I wish we had more time, but we’re at the end of the episode. Thank you so much for joining us—I really enjoyed the conversation.

Allen: Thanks for having me. I appreciate it.

Evan: That was Allen Fazio, Chief Information Officer at Houlihan Lokey.

Saam: Thanks for listening to Enterprise AI Innovators. I’m Saam Motamedi, a general partner at Greylock Partners.

Evan: And I’m Evan Reiser, the founder and CEO of Abnormal AI. Please be sure to subscribe so you never miss an episode. Learn more about enterprise AI transformation at enterprise software.blog. This show is produced by Abnormal Studios. We’ll see you next time.