By Ben Challice, CEO

Where this has been heading

In April, following PASLA Bangkok, we raised a question about how to sequence AI investment. In July, we answered part of it with a working proof point: post-trade behaviour, accumulated across our network, surfacing as a signal at a point where it becomes useful to a trader. Both pieces were building toward one argument, which I will address in this Panorama.

Connectivity used to be the story.

Pirum has spent over a quarter of a century developing automated connectivity for securities lending, fixed income and collateral management. We have partnered with 150+ client institutions, including 25 GSIBs, to build a global network that today moves US$6.5 trillion through it every day, with standardised connections across the securities market.

That scale is hard to copy. Even at a time when AI agents can stand up a product in a fortnight, nobody can replicate 25 years of multi-layered client connections inside a product cycle. But in our view this foundation allows us to offer something more powerful to our clients: intelligence to help make better decisions and provide increased automation.

What the network can now do

The pre-trade, post-trade and collateral behavioural data that this network generates will add value for the desks that produced it. Correctly interrogated, that data can tell a desk where a fail is forming, before it occurs.

It can show how their performance compares with a market they otherwise see only a slice of. And it can move earlier still, to the point before a trade is booked, by flagging a situation to a trader who would not have known to look.

How much it costs to borrow or finance a trade is now well known, but providing the intelligence layer across the lifecycle allows our clients to understand their and their counterparties’ operational efficiency, as well as surfacing operational data points to be considered pre-trade or remediated efficiently, before they become a problem.

Why this year is different

Accenture’s Banking Top Trends 2026 calls this the year agentic AI moves from pilot to production, with human-and-agent teams working the same book in what it calls the “10x bank”.

Google Cloud and SS&C Blue Prism point in a similar direction in their own 2026 research. None of them are writing about securities finance specifically, but the impact should be thought of in the same vein.

Appropriately governed data from a network such as ours can provide the fuel for these agentic AI moves. We are actively working on mechanisms to provide this data to our clients but are also developing agentic solutions for those who will need them across the lifecycle.

Where this goes by 2030

Building on this concept further, some clients envisage Pirum as helping to manage the workflows, such as a managed service, so that they only see the exceptions that require a human in the loop. Others want the signal delivered straight into their own systems – through a real-time API today, or increasingly through MCP – so that their own agents can act on it.

Both are legitimate models and will be waypoints on our roadmap. But what is clear is that, by the beginning of next year, the distinction between ‘the company that connects your trades’ and ‘the company that helps decide how to manage them more efficiently’ will stop being a distinction worth drawing. It will be the same platform doing both.

By 2030, we envisage that the network will act on a meaningful share of its data directly, agent to agent, across counterparties, with human oversight and control retained wherever it adds value or is required by clients’ own governance and applicable regulation. The trader’s job, the job of operations, and the risk manager’s job all start to look more like an orchestrator’s. Rather than relying on distinct pools of data and insight, they will be pulling the various strands together, supported by agents, to deliver additional clarity, intelligence and competitive advantage.

If you would like to speak with Ben Challice about how your firm can move beyond connectivity to intelligence, get in touch. 

Also included in the Quarterly Panorama:

  • Ray Wong, Director of Product, AI & Data, on why there has never been a better time for Pirum to launch AI, and why infrastructure, people and platform must all be ready at once.

  • Zoë Balkwell on pre-trade intelligence, and why the real value of AI is surfacing opportunities a desk would never have known to look for.

  • Renato Lima, Chief Commercial Officer, on the build trap, and why in-house AI starts at near zero when a coding agent cannot replicate 25 years of counterparty trust.

The Panorama also includes a regulatory and events calendar.