Beyond the Hype: Janus Henderson's Blueprint for Embedded AI

SAN FRANCISCO & LONDON – June 11, 2026 – In an industry awash with proclamations about artificial intelligence, asset manager Janus Henderson has made a move that warrants closer inspection, not for its hype, but for its execution-focused strategy. The firm, which oversees nearly half a trillion dollars in assets, announced it is building a suite of AI-native tools in collaboration with Anthropic and Percepta, a transformation company from the portfolio of venture capital firm General Catalyst.

While rivals also explore AI, Janus Henderson’s approach signals a deliberate pivot away from off-the-shelf solutions and toward a deeply integrated model designed to augment, not replace, its human experts. By building custom tools from the ground up and embedding an external AI team directly into its core operations, the firm is placing a significant bet that the true value of AI in asset management lies in bespoke integration, not generic application. This initiative offers a potential blueprint for how legacy institutions can tackle digital transformation, focusing on the practical challenges of merging frontier technology with proprietary workflows.

A New Toolkit for Human Expertise

At the heart of the initiative are two purpose-built platforms, PRISM and LIBROS, powered by Anthropic's Claude AI model. These are not broad, one-size-fits-all applications, but tools engineered to address specific, high-value functions within the firm. PRISM is a global client intelligence platform designed for Janus Henderson's distribution teams. It aims to move beyond basic CRM by using AI to synthesize internal and third-party data, helping client-facing teams identify which clients to contact, understand their needs more deeply, and personalize communications. The goal is to deliver a more consistent and insightful client experience across all regions.

For the investment teams, LIBROS serves as an AI-native research management tool. Asset management is a knowledge-intensive business, and analysts are often inundated with a deluge of internal research, external reports, and public market data. LIBROS is designed to ingest and synthesize this information, helping analysts and portfolio managers surface relevant signals faster. The objective is clear: free up intellectual capital from the drudgery of data collation so that more time can be spent on what active managers are paid for—judgment, critical thinking, and making investment decisions.

"We believe AI Transformation will fundamentally change the way asset managers serve their clients when it is embedded at the core of the business," said Ali Dibadj, Chief Executive Officer of Janus Henderson. His statement underscores the core philosophy here: AI as a force multiplier for human talent. The firm is also deploying Anthropic’s Claude more broadly, with Claude Code for its engineering teams and a general-purpose assistant, Cowork, for employees, signaling a commitment to infusing AI into the firm's operational DNA.

The 'Embedded' Advantage

The most distinctive feature of this collaboration is the methodology. Instead of purchasing a software license or hiring a traditional consulting firm for a high-level strategy, Janus Henderson is adopting Percepta's "embedded model." This involves embedding Percepta's AI engineers, researchers, and product managers directly alongside the firm's own investment, distribution, and technology staff. This hands-on approach is designed to solve a problem that has chronically slowed AI adoption across the financial industry: generic tools rarely fit the nuanced workflows of an active manager.

"The value comes from connecting frontier AI to a firm's own data and rebuilding core workflows around it, which generally takes engineering built into the business, not software bought off a shelf," a source familiar with the project noted. This is the crux of the strategy. By working from the inside, Percepta can construct the data and knowledge foundation that connects Claude to Janus Henderson's proprietary research, client information, and market data—assets that represent the firm's unique intellectual property.

Hirsh Jain, CEO of Percepta, emphasized this point, stating, "Transforming industries with AI requires fundamentally rethinking how work gets done in organizations and engineering systems that are purpose-built for a new mode of operations." This model directly contrasts with the often-frustrating experience of trying to retrofit an external software solution onto complex and sometimes decades-old internal systems. The strategic alignment is further deepened by the fact that Percepta is a General Catalyst company; General Catalyst, along with Trian Partners, signed a definitive agreement in late 2025 to acquire Janus Henderson, making this AI initiative a cornerstone of the firm's future under new ownership.

Claude on Wall Street: A Bet on Reliability

The choice of Anthropic's Claude as the foundational AI model is another critical component of the strategy. In the high-stakes, highly regulated world of finance, attributes like reliability, security, and interpretability are not just desirable—they are essential. Anthropic has built its brand on developing safe and steerable AI, a position that likely resonates with an industry where a single error or biased output can have significant financial and reputational consequences. "Asset management is a knowledge-intensive industry where reliable AI can help teams work faster and serve clients better," said Peter Nolan, Head of Asset & Wealth Management at Anthropic.

Claude's large context window, which allows it to process and reason over vast documents, is particularly well-suited for tasks like synthesizing extensive investment research for LIBROS. However, the path to implementation is filled with technical and ethical hurdles. For these tools to be effective and compliant, they must navigate the "black box" problem, providing enough transparency for users and regulators to understand how conclusions are reached. Furthermore, they must be rigorously tested to mitigate algorithmic biases that could otherwise creep into client recommendations or investment analysis.

While the press release doesn't detail the specific safeguards, the selection of a partner that prioritizes AI safety suggests Janus Henderson is tackling these challenges proactively. For any firm following this path, establishing a robust governance framework around data privacy, model validation, and human oversight is a non-negotiable part of the execution plan.

Navigating the New Frontier

This initiative by Janus Henderson is more than a technology upgrade; it is a test case for the future of active asset management. It acknowledges that in an era of passive investing and fee compression, true differentiation will come from superior insight and service—areas where AI-augmented humans can create a competitive edge. The project's success will depend not only on the sophistication of the technology but on the firm's ability to manage the intricate process of organizational change, reskill its workforce, and satisfy the watchful eyes of regulators like the SEC and FCA.

By choosing a deeply embedded, custom-built path, Janus Henderson is making a calculated wager that the hard work of true integration will yield greater returns than the easy allure of off-the-shelf AI. For industry leaders watching from the sidelines, this grounded approach to innovation provides a compelling look at what it takes to move from pilot to production, turning the promise of AI into a practical business reality.