FNA Sampo: Secure Automated Multi-Party Optimization

Settle payments earlier, with less liquidity and more resilience

Sampo cuts the intraday liquidity banks need by coordinating when they submit large-value payments. Each bank keeps every decision. No netting, no change to the RTGS.

Discuss a pilot How it works

  • Earlier settlement

    2 hours

    earlier on average for each payment, when banks send most of their payment intents before the opening.

  • Less liquidity

    49% less

    peak net debit for the same payments: $55bn across five large banks, simulated.

  • More resilience

    63% less

    value left to the last hour of the day, so an outage or a late payment affects fewer payments.

Sampo Bank view for Bank 1, on synthetic data: liquidity saved, its value and earlier settlement over 20 days, and the advice for the bank's 133 payments.

The Bank view: one bank's advice and liquidity benefit. The evidence

The problem

Banks manage liquidity without knowing when their incoming payments will arrive

A bank knows most of its own payments at the start of the day. Around 70% of large-value payments are known by then. What the bank does not know is when other banks will send the payments it is due to receive.

Without that information, the bank has two options for each outgoing payment: fund it from its own reserves or collateral, or hold it until incoming payments arrive. Banks hold payments and keep buffers against the uncertainty.

The information the bank lacks exists: other banks have it. One bank's outgoing payment is another bank's incoming payment. No bank has all of it, and no bank can act on it alone.

Sampo uses the payments that the banks in the cohort already know, and advises each bank when to submit its own payments, so that incoming payments arrive in time to fund outgoing ones. No bank receives another bank's payment data.

Outgoing payments Known to the bank Incoming payments Timing not known in advance 08:0010:0012:0014:0016:0018:00 ????
One bank's day. It decides when to submit its outgoing payments, but it learns when incoming payments arrive only when they settle.

$4.6tn

settles on Fedwire each day. More than half of that value settles in the final third of the day.

Federal Reserve.

70%

of large-value payments are known at the start of the day.

20%

lower intraday liquidity for CHAPS settlement banks after the liquidity-saving mechanism, about £4bn a day, with no loss of throughput.

Bank of England.

22%

less liquidity from reordering payments alone, on about $1.7 trillion of daily CHIPS payments.

FNA analysis of CHIPS data for The Clearing House.

What it is worth

Calculate the saving for your bank

Enter your monthly peak net debit and your cost of liquidity. The calculator shows the liquidity Sampo releases and its annual value.

What Sampo is

Advice on timing. The bank keeps control over every decision.

Sampo is a compute-and-advice service. It is not part of settlement and holds no funds.

Sampo does

  • Provides advice on submission timing for each bank's payments above an agreed value threshold.
  • Updates the payment schedule every minute, from the payment intents sent so far. Advice is valid for one minute.
  • Returns each bank its own advice only. No bank receives another bank's payment data.

Sampo never

  • Nets. Every payment settles gross and in full, in central bank money.
  • Holds funds or takes part in settlement.
  • Changes a payment's amount, sender or receiver.
  • Asks for balances or liquidity positions.
  • Enforces. Sampo advises. Each bank submits its own payments, through its own systems.

For central banks and supervisors

Earlier settlement, with less value left until the end of the day

Liquidity-saving mechanisms are treated as financial stability infrastructure as well as efficiency tools. FNA shows with Sampo that a cohort of banks can obtain that benefit without a change to the RTGS.

Financial stability

When banks hold payments until late in the day, value concentrates in a short window and queues build. An outage, a late payment or a stressed bank then affects more payments. Earlier, steadier settlement reduces that exposure.

Efficient use of reserves and collateral

Reserves held for intraday payments have an opportunity cost, and collateral pledged for intraday credit cannot be used elsewhere. Lower intraday peaks mean the same payments settle with less of both.

No change to the RTGS

Sampo is advisory. The central bank keeps the RTGS as it is, settlement stays final in central bank money, and oversight roles do not change.

Request a briefing

Sampo views

Each party receives only the data its role allows

A bank receives its own advice and benefit figures, as in the Bank view above. Monitoring receives system health data without amounts. The Simulation view shows the outcome for the whole cohort. All three screens show results from a synthetic five-bank model.

Monitoring view on synthetic data: 500 large-value payments, 131 submitted at their own time, 343 payment schedules sent, and an intraday timeline of submissions. No amounts are shown.
System health, submissions and updates of the payment schedule, with no payment amounts.

How it works

Six steps, from payment intent to measured saving

Below: one model day, replayed. In Sampo's design, independent compute parties compute the payment schedule together, and each sees only shares of the amounts.

Diagram of Sampo's design, with banks, three compute parties and monitoring. Each bank sends payment intents to the compute parties. They compute the payment schedule together and send each bank its advice.

  1. Bank sends payment intents

    Each bank sends Sampo a payment intent for each of its upcoming large-value payments. A payment intent is not a payment instruction.

  2. Sampo computes the payment schedule

    Sampo computes a payment schedule with Sampo's rule, parts of which are patent pending. Today Sampo runs on payment data in the clear. In Sampo's design, independent compute parties compute the payment schedule together, and each sees only shares of the amounts.

  3. Sampo sends each bank its own advice

    Sampo sends each bank advice for its own payments only.

  4. Bank submits payments; the RTGS settles them

    The bank submits each payment to the RTGS when Sampo advises it, or at its own time, using its existing systems. The RTGS settles it as today.

  5. Sampo updates the payment schedule every minute

    Sampo updates the payment schedule every minute from the payment intents sent so far and advises which payments to submit now. Advice is valid for one minute. The bank submits a payment that Sampo does not advise early at its own time.

  6. Sampo monitors submissions and reports the saving

    Sampo compares actual submission times with its advice and reports the liquidity saved. It uses submission and receipt times, not amounts.

The operating flow and privacy model in full

Work with Sampo

Measure it on your own flows, or test it with your own agents

For banks, central banks and RTGS operators

An evidence pilot on your own payments

We replay your agreed historical or shadow flows and compare your bank's peak without Sampo with its peak with Sampo. You get your own bank's figure, not a network average.

Discuss a pilot

For researchers and agent developers

The AI Sandbox

AI agents act as banks in a simulated payment day. Each agent sends its payment intents to Sampo, receives Sampo's advice and submits to a simulated RTGS. Agents connect over the Model Context Protocol (MCP) or an HTTP API. Synthetic data only.

About the AI Sandbox

Why FNA

We build infrastructure that banks and authorities use together

National infrastructure in production

FNA built and operates the world's first national anti-scam utility: Malaysia's National Scam Response Centre, live across 48 institutions.

Tested on real large-value data

FNA's review of CHIPS data for The Clearing House measured the opportunity from reordering payments on about $1.7 trillion of daily payments.

Work with FMIs and regulators

FNA works with RTGS operators, financial market infrastructures and regulators worldwide.

Privacy by design

In Sampo's design, independent compute parties compute the payment schedule together, and each sees only shares of the amounts. Payment amounts would then stay confidential from the service operator, every compute party and every other bank.

Selected FNA project experience

FNA has delivered analytics and simulation projects with central banks, payment-system operators and industry bodies.

  • Bank of England
  • The Clearing House
  • UK Finance
  • CLS
  • Monetary Authority of Singapore
  • Payments Canada

BIS Innovation Hub and MAS

Project Ellipse

FNA supported the BIS Innovation Hub Singapore Centre and the Monetary Authority of Singapore, with the Bank of England and others, on a prototype platform for supervisory data.

BIS project report

The Clearing House

Independent CHIPS review

FNA assessed CHIPS' liquidity efficiency using payment data and simulations of alternative settlement arrangements.

Project announcement

UK Finance

CHAPS liquidity study

FNA modelled how pre-opening payment compression could lower intraday liquidity for CHAPS banks.

Read the case study

FNA's wider collaborations include the Central Bank of the UAE, Saudi Central Bank and SWIFT.

Research

Research on coordinating payment timing

Morten Bech and Kimmo Soramäki formalised gridlock resolution in interbank payment systems in 2001. That work describes the class of mechanism that saves liquidity in CHAPS today.

Briefing

The intraday liquidity problem

FNA · 2026

Why liquidity use is uncoordinated, what regulation requires, and why coordination also improves resilience.

Read
Briefing

Data and methodology

FNA · 2026

How the modeled results are produced and what they leave out.

Read
FNA co-author

Gridlock resolution in interbank payment systems

M. Bech and K. Soramäki · Bank of Finland · 2001

The paper that formalised how to resolve gridlock in RTGS payment queues.

Source

The research library: FNA research, briefing papers and the literature

Where Sampo stands

FNA is selecting design-partner banks

Sampo is entering a USD shadow-data pilot on real bank flows. Design partners measure the saving on their own payments first.

Become a design partner Request access

Or email Phillip Straley at phillip@fna.fi