Asset owners sit on more data than ever, spread across disconnected pockets, and multiplying AI agents make it harder, not easier, to control. How can a domain-aware control plane orchestrate data and intelligence for better investment decisions?
Investment data has never been more plentiful, or more scattered. Familiar data management challenges arise from multi-asset portfolios, with each asset class bringing their own requirements, while ESG commitments and new regulation add further demands.
Asset owners face a newer challenge: orchestration. Think of the operating model as an orchestra and data as its sound. It flows in from every direction: managers, vendors, ESG providers and administrators, each playing in its own format and language. It flows out again to investment teams, risk, reporting and regulators, who all need to hear it clearly.
With numerous AI agents now joining the orchestra, intelligence sprawl is joining data sprawl, and asset owners know their data is not ready. A recent Thinking Ahead Institute and MSCI Institute study found that around three-quarters cite data quality and standardisation as a challenge in keeping pace with AI, more than any other issue.Two-thirds say poor integration constrains the quality of their investment decisions.
The challenge is not deploying AI. It is ensuring that multiple AI-driven processes operate consistently, transparently and from the same understanding of investment data. Without a conductor, what comes in and goes out is noise. With one orchestrating both the data and the intelligence, every player works from the same score, and the result is harmony: better investment decisions.
Investment data is an operating model problem, not a technology one
The instinctive response to complexity has been to add technology: another platform, another database, another dashboard. The real problem is an operating model that lacks control. First, it was data silos; now it is orchestration sprawl, with every tool, feed and agent integrated in its own way. What is missing is industrialisation: treating investment data as a production process, with consistent standards, controls and ownership, rather than a series of fixes.
While AI can tackle fragmented data, improve data quality and ease reconciliation, it can only create value if the data is properly processed and can be trusted. Gartner® predicts that “through 2026 firms will abandon 60% of AI projects that are not supported by AI-ready data.” [1] A secured, trustworthy investment data control plane becomes crucial for firms to progress with any meaningful AI projects.
What goes wrong when agents don’t talk to each other
Consider an asset owner running an ESG agent that handles proxy voting and a risk agent that produces trade recommendations. Neither communicates with the other. A vote or a trade may not fit the investment mandate, and the fund could breach a limit without anyone noticing. The fix might be yet another agent to check compliance, adding to the sprawl. Because AI works far faster than any manual process, errors compound just as fast.
What is a domain-aware investment data control plane?
A control plane is not a new idea, but it has a new job. It now needs to be an orchestration layer not just for data, but for intelligence itself. An investment data control plane is the layer that every piece of data, and every AI agent, passes through. Data enters it, has controls applied and is routed where it needs to go. Agents are governed in the same way: how they talk to each other and to the asset owner’s systems, what they can act on, and which exceptions go to people, because human judgement remains part of the control. It does not replace a database or warehouse; it controls what goes into them.
Domain awareness is what makes the difference, and it depends on people. AI is only as good as the hands directing it. Domain awareness puts human experts in the loop: people who understand the business of investing, know how data connects across mandates, managers and asset classes, and can control the AI rather than defer to it. That is the core of a managed service. It brings together specialists in investment, operations and data, a mix that is rare and expensive to hire in-house, to classify, enrich and link data so that it reflects how the fund actually invests.
From pockets of data to a total portfolio view
Industrialising investment data is what turns those pockets into one overview. It means acquiring, normalising and enriching data from every source, with the same rules and controls applied. The result is a total portfolio view of every asset, whatever its location, type or manager, that can be aggregated across public and private markets and drilled down to individual manager positions. This is also what makes AI usable. An agent working from the total portfolio view sees the same exposures, limits and mandates as every other agent and every person on the team. Industrialisation gives AI a shared understanding of investment data; the domain-aware control plane, which delivers it, also makes sure every agent acts on it consistently and transparently.
Operating model choices: open architecture or single provider?
Asset owners often frame the choice as open architecture versus a single provider. The real question is the integrity of control. A fund with multiple managers, service providers and accounting systems needs a control plane that orchestrates across them and produces a golden record.
Why the custodian is well placed
Asset owners are increasingly asking service providers for a clear recommendation, not just neutral support. The custodian already sits between the asset owner and much of this traffic. It holds the book of record, applies governance and controls every day, and has an overview of the fund’s assets and investment data that no single manager or vendor has. It also knows how comparable asset owners have tackled the same problems.
From this position, the custodian can stay agnostic about the clients’ tools while being clear about one thing: the need for a domain-aware control layer in the operating model.
The Thinking Ahead Institute and MSCI Institute study found that few asset owners have the scale, expertise and capacity to overhaul their data architecture alone, and building a control plane in-house is costly [2]. The cost is linear: every new challenge brings another point solution and another silo.
A managed, cloud-based service, provided by a custodian that understands the investment business, can best support the foundation of an investment data control plane. BNP Paribas Securities Services delivers this through its Data PRISM360 solution, in partnership with NeoXam. Because it’s light touch and can work with existing tools, it can be in place faster and without the cost and disruption of a full build or multi-year programme. The asset owner keeps control of its own orchestra, while its operational and investments teams get the systems and expertise to focus on the decisions that matter.
From data access to data industrialisation
The future competitive advantage for asset owners is not data access. It is data industrialisation. The winners will manage investment data as a utility, with the domain-aware control plane as the foundation for AI, analytics, reporting, performance and risk oversight. The first step is a change of mindset. Asset owners are no longer just cleaning data; they are managing it, and increasingly managing intelligence too. With orchestration under control, data and intelligence work in harmony: AI the fund can trust, and better investment decisions.
- Thinking Ahead Institute and MSCI Institute, Global Asset Owner Technology & AI Peer Study 2026, October 2026, https://www.thinkingaheadinstitute.org/research-papers/global-asset-owner-technology-ai-peer-study/
- Gartner, “Lack of AI-Ready Data Puts AI Projects at Risk”, press release, 26 February 2025, https://www.gartner.com/en/newsroom/press-releases/2025-02-26-lack-of-ai-ready-data-puts-ai-projects-at-risk
GARTNER is a trademark of Gartner, Inc. and/or its affiliate
For more on the investment data management challenges asset owners face
