In securities lending, the honest answer is uncomfortable. The data many firms would feed these tools were emailed, faxed, re-keyed, and re-requested several times before the first loan was ever booked. The industry has spent two decades automating what happens after a loan is struck. The front door of the business, the onboarding and maintenance of the lending accounts themselves, still runs on processes that predate the smartphone.
The gap nobody books
Before an account can lend a single share, its details, legal agreements, tax documentation, and market requirements pass through onboarding, legal, tax, credit risk, and operations. Each department often re-requests what another has already collected. And the problem compounds as the industry reaches for growth, because every new lending market begins life as another set of accounts to open.
That gap has a cost that rarely gets booked. Every week between a signed mandate and the first loan is lending revenue the beneficial owner may not earn and inventory the agent lender may not mobilise. Ask firms how long a new lending account takes to activate and the honest answer is “it depends”, because no one has a live view of what remains outstanding. The delay may rarely appear as a loss on any ledger. It shows up as revenue that arrived late or not at all, which may be why it has survived this long.
What changed this year
A different model has moved from argument to adoption. Digitise account data and documents once, at the point of setup, and maintain memory of them for the life of the account. Lender, agent, and approved borrowers work from the same consistent data from day one. Documentation packages arrive pre-populated. Updates flow to all permissioned parties in near real-time. And every authorised party sees exactly what stands between an account and its first loan, a live ready-to-lend status per account.
This is no longer a future-state concept. Several leading agent lenders and direct lenders are now moving their lending account onboarding and maintenance onto this model, for a simple reason. The faster an account is ready-to-lend, the sooner it can start earning.
“Securities lending often begins with an account already established for custody or other trading activity,” says Gabino Roche Jr, CEO and founder of Saphyre. “The opportunity is to intelligently reuse the data and documents already provided, then keep that information current throughout the account’s life. Shared market infrastructure connecting clients, custodians, lenders, and borrowers makes that possible across permissioned counterparties; helping firms start lending sooner and keep accounts ready for ongoing business.”
Where the model pays twice
The economics improve further for firms that wear more than one hat, which in this industry is many of them. An agent lender that also custodies assets can run a shared onboarding that serves both custody and lending, because the legal entity data, tax documentation, and settlement instructions are largely the same. A direct lender that also manages assets can reuse the fund onboarding, Know Your Client (KYC), and tax information already collected on the asset management side when facing borrower counterparties, rather than requesting it again for the lending programme.
In both cases, the second use of the data costs far less than the first. That is the advantage of maintaining memory rather than collecting documents transaction by transaction.
Agency Lending Disclosure (ALD) shows why that memory matters beyond onboarding. Borrowers need current information about the underlying principals from whom they borrow, yet requests can send agent lenders back to clients for documents and details already supplied.
Retaining the links between each principal, its accounts, supporting documents, and applicable agreement terms makes existing information easier to retrieve and share with permissioned parties, while allowing new information and updates to be incorporated as the relationship evolves.
As principals join a lending programme or further loans are made from existing accounts, that context supports borrower approvals and helps firms identify the loan terms and collateral requirements applicable under the relevant Master Securities Lending Agreement (MSLA) or Global Master Securities Lending Agreement (GMSLA). Alongside ongoing disclosure of loan and collateral allocations, this continuity supports exposure monitoring and reduces repeated information gathering throughout the lending lifecycle.
One year to T+1 Europe, and the standard setters have weighed in
The regulatory calendar gives this argument an edge this week. The UK, EU, and Switzerland move to T+1 settlement on 11 October 2027, almost exactly one year from this conference.
The industry’s standard setters have already pointed to where part of the fix lives. The Financial Markets Standards Board’s (FMSB’s) Standard for Sharing of Standard Settlement Instructions, published in January 2025, states in its third Core Principle that standard settlement instructions (SSIs) expected to be used with a counterparty should be shared at the point of onboarding, covering all products referenced in the legal agreements, which includes securities lending.
The UK’s Accelerated Settlement Taskforce wrote both ideas into its implementation plan. It expects all market participants to implement the FMSB’s Core Principles, and it calls on lending intermediaries and borrowers to automate recalls processing through standardised electronic messaging. The EU’s T+1 roadmap makes the same diagnosis, describing SSIs as “an inherent weak point in post-trade processing” and putting automation of securities lending recalls and returns among its recommendations.
Read together, the message to securities lending is hard to miss. Many of the fixes the taskforces want live in pre-trade, at onboarding, in the data. Many fails are born months before settlement, at account setup.
New tools, old data
Which brings the conversation back to the week’s headline debate about what will run the next market. The instinct across the industry right now is to point new technology at the mess, and the mess is real. But AI models trained on inconsistent counterparty data do not resolve the inconsistency. They automate it, at scale, with confidence. A more connected market hits the same wall, because platforms cannot connect on data that disagrees. Whatever wins the debate will only be as good as the account data underneath it.
“AI is giving our industry a false sense of confidence,” says Stephen Roche, president and co-founder of Saphyre. “If the underlying data doesn’t agree across your counterparties, AI just automates the disagreement faster. Get the data consistent at the point of setup and AI finally has something reliable to decide on.”
The takeaway
Onboarding used to be seen as part of the cost of doing business. It is now a revenue enablement advantage. Firms that treat lending account setup as a digital, data-first discipline can activate relationships faster, lower operating costs on both sides, strengthen ALD from inception, and carry data consistency through settlement. They are also building the one thing every innovation on this week’s agenda actually requires, and they will be a year ahead when T+1 arrives in Europe.
The question worth asking as this conference gets underway is the same one worth asking of every account in your programme. Do you know, today, exactly what stands between it and ready-to-lend?
← Previous interview
Bank of America
Insights and perspectives
Image: Saphyre 
.jpg)