On the 61st episode of Enterprise AI Innovators, hosts Evan Reiser (CEO & co-founder, Abnormal AI) and Saam Motamedi (Greylock Partners) talk with Allen Fazio, CIO at Houlihan Lokey, one of the global leaders in mid-cap M&A, debt restructuring, and valuation. Allen’s background spans public accounting, two decades inside Disney’s theme parks and cruise lines, and CIO roles at Disneyland and Walt Disney Imagineering. His move into investment banking placed him at the center of a firm that grew from roughly $700 million to $2.5 billion in revenue. That broad career arc shapes how he thinks about AI as infrastructure for knowledge work rather than a novelty or a chatbot.
At Houlihan Lokey, AI is not emerging gradually. It is arriving quickly. Four months ago, the firm detected around 60 AI-branded tools across its network. The most recent scan counted 190, excluding AI within existing software. For a professional services-style firm that depends on analysts to produce models, commentary, and research, that swelling toolset creates both opportunity and operational strain. Allen and his team have spent nearly three years searching for what he calls the orchestration layer. As he explains, “We have been looking for almost three years for what I will call the orchestration layer. What do we put in the middle of the picture?” His goal is to train one model on Houlihan Lokey’s twenty-year trove of M&A data and build custom agents and workflows around that unified system.
Allen believes strongly that AI should enhance human expertise. As he puts it, “We believe a banker with AI is better than a banker. We do not believe AI is better than a banker with AI.” This belief informs how the firm thinks about the future of talent. Entry-level service desk analysts already benefit from a significantly stronger knowledge base. In investment banking, similar questions arise. If AI can extract crucial elements from a 900-page contract in minutes, what becomes the new path for developing the judgment that senior bankers rely on? Allen acknowledges the uncertainty but emphasizes that both skills and roles will need to evolve.
Usage data inside Houlihan Lokey makes the adoption pattern clear. Analysts lead in AI usage, followed by associates, VPs, and directors. When interns arrived, their usage lines quickly overtook those of associates and approached those of analysts. Instead of resisting that trend, Allen is designing around it. “Our focus is going to be on the lookout level,” he says. He is centering the early phase of AI adoption on interns, analysts, and associates through 100-level work such as drafting, summarization, and research. Higher-level 400-level use cases are being added more slowly to avoid overwhelming senior bankers or creating early negative experiences that undermine trust.
Cultural adoption matters as much as technical rollout. One example involves a tool that summarizes a week of meetings and emails in the style of a Comedy Central roast. According to Allen, “Every Friday we say, could you please summarize my week in the form of a Comedy Central roast, and they are brilliant.” What began as a humorous experiment has turned into a practical workflow for tracking commitments and follow-ups. For Allen, this illustrates how a culture of innovation spreads through enjoyable and helpful entry points rather than through top-down mandates.
Regulation adds complexity that is unique to investment banking. Conversations that historically vanished once a meeting ended now become digital artifacts once AI-driven transcription tools are involved. Allen raises questions about whether firms will need to store raw recordings, transcripts, or only summaries, and how prompts influence the accuracy and completeness of those summaries. These issues guide his work with legal partners and reinforce the need for deliberate sequencing of AI adoption.
Looking ahead, Allen believes future technology leaders must translate technical ideas into clear business language. His own reading spans leadership texts, technology books, and science fiction such as Snow Crash, which he is currently revisiting. He predicts that the next generation of analysts and technologists will stand out not through perfect spreadsheets but by mastering AI tools, validating results, and delivering higher-quality answers faster.
Listen to Allen’s episode here and read the transcript here.
