Independent analysis of AI for payments and finance operations

We help payments and finance leaders decide where and how to use AI, using model tests to work out the economics.

Research Advisory

Free to use. Sign up to get updates.

Our research agenda

We test AI models on real payments and finance tasks and score them on the economics: what their mistakes cost and what they cost to run. Each test takes one task from a map of 51 payments and finance processes and answers three questions about each model. How the tests work →

1

How well does each model do a given payments task, and what do its mistakes cost?

Each test has a leaderboard that ranks models on the money they get wrong.

2

What does each model cost per task?

Every leaderboard shows the model cost per task next to the score.

3

Where can each model run with zero data retention?

The hosting page shows the route each model ran on and whether that route keeps a copy of the data.

What the tests show so far

Every chart below comes from the live leaderboards and updates when a test does. Enter research for every result.

Reconciling card and payout income See the leaderboard → Reading card statements See the leaderboard → Rankings across the tests Score out of 100. Teal models have taken both tests. See the rankings → Open weights against closed Score out of 100 on each test, best model of each kind. Open-weight models can run on a firm's own servers. See hosting and terms → Where the tests sit Two of the 11 clusters of payments and finance processes have a model test so far. See the process map →

From test result to economic model

Each test has an Excel model, available on request, that turns the results into cost per task, annual cost and payback for a given firm.

See the Test 2 model → A model for your firm →

Free to use. Get an email when a new test is out.

Sign up to get an email when a new test or price update comes out.

Articles

Test 1

Reading Card Statements

What Test 1 found when 34 models read UK card acquiring statements, which faulty statements they caught, and what each model costs to run.Read → Research · Model tests

Test Design: What the Model Does and What Code Does

Which steps the tests give to a model, which they leave to code, and the judgement skills they score.Read → Latest test · Test 2

Reconciling Settlements

What Test 2 found when 11 models reconciled a month of card and payout settlements, and how far apart the models were.Read → AI economics

The Unit Economics of AI in Payments and Finance Operations

What an AI model really costs per case once checking, mistakes and fixed costs are counted.Read →
All articles →

Custom research and advisory

The wider agenda. Beyond the tests, ScalePoint Research covers investment prioritisation, budgeting, forecasting, cost management and value realisation for AI in payments and finance operations.

  • Economic models. A model of a use case's cost structure, unit economics, payback and scenarios, with one-off investment kept apart from recurring costs and forecast benefits kept apart from measured ones.
  • Build, buy, combine, scale or stop. Whether to build, buy or combine for a given process, which deployment option fits, how to fund it, and when to scale, redesign or stop a use case.
  • Model and Vendor Check. Private evaluations of AI models, or a vendor's AI product, on your own cases: accuracy in money, cost per task and where failures sit, plus acceptance tests for the contract.
  • Pricing intelligence. Price points for models, hosting and vendor products, and the commercial terms that change cost at scale, built into the economic models.
  • Value realisation. How adoption, workflow change, output quality, human oversight and risk affect the cost and the value actually delivered, checked against the business case for live use cases.

About

Martin Koderisch

Martin Koderisch is the founder and lead analyst of ScalePoint Research. He leads the research on the economics of AI in payments and finance operations, and advises payments and fintech businesses on strategic finance, unit economics and value creation.

He spent ten years in operating roles at Citi, Mastercard and the start-up Luup, then nine years at Edgar, Dunn & Company, latterly as a Principal, on more than 75 engagements for banks, fintechs, card networks, private equity firms and merchants. Since 2024 he has worked independently with payments and fintech businesses on strategic finance and value creation. He splits his time between London and Barcelona.

Email martin@scalepointpartners.com or connect on LinkedIn.