The Constraint On Research Modernization is Not Technology

Why research franchises that modernize client engagement early could compound advantage for years

The most commercially important layer of investment banking remains one of its least modernized. Banks spend heavily on technology, but relatively little has transformed the client-facing side of the business. Even in the AI era, adoption has concentrated in lower-risk functions such as compliance and KYC, while the layer where revenue share is won or lost still operates through fragmented workflows and legacy systems.

Nearly every major industry in the world has built systems to anticipate customer needs and reduce friction—not out of generosity, but because ease of engagement drives revenue. In institutional finance, by contrast, buy-side clients still encounter surprisingly basic obstacles: finding relevant research from a trusted analyst, discovering upcoming events, or navigating who to contact and when.

The capability to do much better already exists. The belief that it matters enough to prioritize still does not.

“But finance is different.” The objection is familiar: institutional finance is too regulated, too relationship-driven, and too operationally complex and exceptional for the kinds of commercial disciplines that drive growth elsewhere—targeted engagement, strategic marketing, and lower-friction client experiences. Meanwhile, incumbents in adjacent industries have repeatedly overestimated how durable those barriers are.

Pharma offers a useful parallel. It operates in a similarly regulated and relationship-driven environment, yet leading firms have reported meaningful returns from AI-driven physician engagement, with published case studies showing double-digit sales lifts among representatives acting on AI-generated targeting.

Perceived uniqueness rarely protects incumbents for as long as they expect it to.

What does AI-powered research engagement actually look like? Imagine the experience of a buy-side portfolio manager. Today, she receives 300 sell-side emails a week, most of which she will never open.

Now imagine the same experience with AI-enabled research engagement. Her morning digest surfaces the handful of notes most relevant to her portfolio and recent activity, while everything else is condensed into a concise brief. As she reads research on a company, the system surfaces an upcoming conference where management will be attending, allowing her to register interest in real time. As her behavior changes, so does the system. Repeated engagement with a sector elevates relevant research, analyst access, and upcoming management meetings. Follow-up responses after calls are drafted with supporting context and prior research attached, then routed through the analyst for refinement. To the client, the experience still feels personal. To the bank, it becomes personalized at scale.

None of this is theoretical. Variations already exist in adjacent industries. What is missing is not feasibility, but conviction that this matters commercially.

Why does this matter now? Research economics are stabilizing after years of pressure. The U.S. equities commission wallet rose to nearly $6.2 billion in 2024 and grew another ~12% in 2025 to just under $7 billion. At the same time, budgets are concentrating, with roughly 10 firms capturing ~55% of research spend.

That concentration creates room for meaningful share gains outside the top tier. Consider Bank Y: a credible franchise at ~2.5% wallet share with solid coverage and established client relationships. An additional ~150bps of share would likely push it into the top 10—ambitious, but not unprecedented. Jefferies climbed from #10 to #4 between 2021 and 2025 through the traditional playbook of strong hiring, expanded coverage, and countercyclical investment.

AI does not change that playbook, but it compresses the time it takes for it to work. What was previously operationally difficult at scale—personalized engagement, rapid feedback loops, and real-time client signaling—becomes continuously executable. Senior hire impact, client engagement, and research consumption become visible in near real time, allowing coverage, access, and outreach to adjust continuously rather than retrospectively.

In that environment, the same strategy produces faster and more durable share gains. Bank Y’s share rises ~100–150bps or more, particularly when paired with a catalyst such as an acquisition, a regulatory opening, or sustained multi-year talent build. Just as importantly, gains become stickier as engagement histories deepen and compound. As I mentioned in a related blog post, over time, the research franchise becomes embedded in the intellectual operating system of the client, and it’s that attention and influence that generates commercial value.

AI should change the work, not replace the worker. Much of the anxiety around AI in institutional finance centers on replacement risk. In practice, its near-term impact is less about replacement and more about reallocating time toward higher-value work.

AI is already effective at automating administrative tasks that have consumed disproportionate senior time for years: formatting materials, cleaning data, drafting routine communications, and compiling reports. If that is the bulk of someone’s value, concern is warranted. But the real value of senior roles lies in judgment, relationships, interpretation, and insight—capabilities that become more powerful when time is redirected toward them.

The constraint is not technology, but its application. The advent of AI has made personalization and frictionless engagement at scale entirely possible. The question is whether firms believe it will drive revenue, and therefore whether they are willing to commit to building it.

The cost of AI itself has fallen dramatically and is now the inexpensive layer. The primary cost lies in infrastructure: data acquisition and cleanup, systems integration, and operational execution. While these are substantial investments of both time and money, they are also compounding investments.

Engagement histories deepen. Recommendation systems improve. Workflow integrations become harder to replicate. The gap between firms with integrated client intelligence systems and those without them widens over time. The firms that move first will not just gain advantage—they will set the expectations others are forced to follow.

The question is who will act first before expectations reset around them, and who will be left behind with a prolonged recovery ahead.

As always, would love to discuss.

KTB