Janus Henderson, a global active asset manager, says it is taking a more embedded approach to enterprise AI by developing AI-native tools for investment research and client-facing teams, powered by Anthropic’s Claude. The firm’s plan, carried out with Percepta, a General Catalyst transformation company, highlights a broader trend in asset management: moving beyond generic AI chat interfaces toward systems connected to internal data, workflows, and day-to-day responsibilities.
What Janus Henderson is building
According to the announcement, Janus Henderson will use Claude as the AI model layer while Percepta supports the infrastructure and the integration work required to connect the system to the firm’s existing knowledge and data environments.
PRISM for client intelligence and engagement
One of the new products is PRISM, described as a “global client intelligence and engagement platform” for Janus Henderson’s distribution teams. The company says PRISM is designed to help client-facing staff prioritize outreach, use internal and third-party data to understand what clients hold and need, and prepare personalized communications.
The initiative is positioned as a unified tool across regions for sales and marketing teams, reflecting an operational focus on how institutions tailor communication at scale while maintaining consistency across markets.
LIBROS for research management
For the investment organization, Janus Henderson is developing LIBROS, an AI-native research management tool for analysts and portfolio managers. The firm says LIBROS will synthesize Janus Henderson’s internal research, external research, and public market data to help surface relevant signals more quickly.
In practical terms, the company is aiming to compress parts of the research workflow that often involve finding, summarizing, and comparing multiple sources, so that more time can be spent on judgment and investment decisions.
Deploying Claude across teams
Janus Henderson also says it is deploying Claude more broadly across the organization, beyond PRISM and LIBROS. The firm references two additional internal uses: Claude Code for engineering teams and Cowork for employees across investment, distribution, and corporate functions.
While the announcement does not provide performance metrics or timelines for each rollout stage, the inclusion of engineering and workplace tools suggests the company is treating AI as an operational layer that can affect both software development and internal productivity, not only client communications and research.
Why the enterprise integration matters
A central point in the release is that generic AI tools often do not map cleanly to the way active asset managers conduct research, manage portfolios, and support clients. Janus Henderson attributes the value to connecting advanced AI to the firm’s proprietary research, client, and market data, and to rebuilding workflows around that connection.
This is a recurring challenge in enterprise AI adoption. Model quality is only one factor, and institutions typically need additional capabilities, including data access controls, knowledge management, workflow redesign, and governance to ensure outputs are usable and safe for professional contexts.
Percepta’s role is described as embedded support, with teams working alongside Janus Henderson staff to build the “data and knowledge foundation” needed to connect Claude to the firm’s internal materials. Percepta also references the use of its Mosaic platform to deliver agentic workflows and decision-support tools, as part of what it says is a broader enterprise transformation playbook.
Industry implications for AI-native asset management
Janus Henderson’s announcement fits within a wider shift across financial services, where large firms are increasingly experimenting with AI but are focused on integration rather than standalone applications. In asset management, the stakes are particularly high because research processes rely on complex knowledge, multiple data sources, and review cycles, while client-facing activities require both personalization and operational control.
From a business standpoint, AI-native tools like PRISM and LIBROS could influence:
- Distribution productivity: improvements to how client outreach is prioritized and tailored may reduce manual steps for sales and marketing teams.
- Research workflow speed: faster synthesis of internal and external materials can change how analysts allocate time between information gathering and decision-making.
- Consistency across regions: a single engagement platform can standardize processes while allowing for data-driven personalization.
- Controls and governance needs: enterprise rollouts typically require careful design to manage data access, output reliability, and internal approval processes.
However, the announcement does not disclose implementation costs, expected benefits, or how the tools will be evaluated over time. As with most AI deployments, outcomes will likely depend on data quality, model governance, user training, and the extent to which workflows are truly redesigned around the tools.
Partnership landscape
The collaboration brings together Janus Henderson, Anthropic, and Percepta, each positioned with a role in the stack: Claude as the model layer, Percepta as the transformation and integration partner, and Janus Henderson as the operator building the application workflows that reflect its active management approach.
The news comes as asset managers continue exploring how AI can support knowledge-intensive tasks, from research to investor communications. For the industry, the key question is whether these “AI-native” tools can deliver repeatable operational advantages without introducing unacceptable compliance and risk overhead.
Note: The companies’ statements in the release include forward-looking language, but specific performance outcomes were not included.
