The best way to use PayPal for business
Discover the key advantages of a PayPal business account – and avoid costly forced currency conversions with amnis' multi-currency acquiring solution.
Most forecasts about SME spending are based on surveys. Ours are based on what businesses actually do. Looking across our platform, some clear trends have emerged, and they paint a striking picture: AI adoption among our SME customers has doubled in the last 12 months.
The numbers
As of March 2026, nearly 1 in 3 amnis SME customers is actively paying for a leading AI platform:
-
- 15.98% of our SME customers hold an active OpenAI or ChatGPT subscription
- 11.4% hold an active Anthropic or Claude.ai subscription
These aren’t vanity metrics or survey responses — these are real transactions processed through amnis. SMEs are putting real money into AI, and the pace is accelerating.
A year ago, these numbers were half what they are today. That kind of growth doesn’t happen by accident. It happens when a technology crosses from “interesting experiment” to “operational necessity.”
What this means for SME banking
The SMEs investing in AI aren’t doing it to impress anyone. They’re doing it because it works — because it helps them do more with the teams they have, move faster than their competitors, and focus human attention where it actually matters.
But here’s the gap: most business banking infrastructure wasn’t built for this. It was built for a world where financial operations were inherently manual — import a file, check a field, approve a payment. That world is ending.
That gap is closing. And we think it closes from the banking side, not just the AI side.
What we’re building at amnis
We’re not observing this shift from the sidelines. We’ve been building for it.
Two AI-powered features are already live on the amnis platform today:
-
- Expense AI automates the categorisation, review, and processing of expenses — turning what used to be a manual, error-prone workflow into something that largely handles itself.
- Payment AI extracts payment data directly from documents — invoices, remittance advice, purchase orders — and initiates payments without requiring manual data entry. Upload a document, confirm, done. You can also use Mail-to-Pay: simply forward an invoice to your dedicated amnis email address, and amnis automatically processes it and creates a payment draft as a complementary workflow.
These aren’t bolt-on features. They’re the first expression of a larger architectural shift: a platform where AI handles the operational layer so finance teams can focus on the strategic one.
amnis AI Roadmap
AI-powered banking — live & coming soon
Live now In pipeline
Live
Expense AI
Automated categorisation, review, and processing of expenses — no manual input required.
Live
Payment AI
Extract payment data from invoices and documents. Upload, confirm, done.
Pipeline
Conversational Banking
Query balances, transactions, and accounts in plain language — no dashboards needed.
Pipeline
AI Analytics
Surface patterns, anomalies, and financial insights automatically from your data.
Pipeline
Treasury Agent
Monitors liquidity, flags risks, and executes treasury actions within set parameters.
Pipeline
Expense Agent
End-to-end expense management running in the background — from receipt to reconciliation.
Pipeline
Subscription Savings Agent
Identifies redundant or underused subscriptions and executes consolidations.
Pipeline
Accounting Agent
Autonomous transactional accounting — classifies, books, and reconciles entries without manual input.
What’s next
The roadmap ahead is where it gets interesting. Here are some areas we’re exploring as we think about what a truly agentic experience for treasury and finance teams could look like.
-
- Conversational banking — interact with your accounts, balances, and transactions in plain language
- AI analytics — surface patterns, anomalies, and insights from your financial data without building reports
- Subscription savings agent — automatically identify redundant or underused subscriptions and propose consolidations
- Treasury agent — an autonomous agent that monitors liquidity, flags risks, and executes treasury actions within defined parameters
- Expense agent — end-to-end expense management that runs in the background, from receipt capture and reminders to reconciliation
- Accounting agent — autonomous transactional accounting that classifies, books, and reconciles entries without manual input
Each of these is designed around the same principle: your finance team’s time is valuable. The routine work should be automated. The judgment calls should stay with humans.
MCP server — beta access
The data makes one thing clear: your teams are already working in AI tools. The question isn’t whether AI will be part of your workflow, it’s whether your banking platform can keep up with it.
We think it should. Finance shouldn’t stop at the edge of your banking dashboard. It should be accessible wherever your team actually works. That’s why we’re building an MCP (Model Context Protocol) server for amnis — an open standard that lets AI assistants like Claude and ChatGPT connect directly to external systems. In practice, it means your AI tools will be able to query your amnis account, pull live balances, initiate payments, and surface FX positions — without switching context, exporting files, or copying data across platforms.
Ask your AI assistant “what’s our EUR cash position across all accounts?” and get a live answer, straight from your banking data.
Sounds interesting? We’re looking for small group of customers and early-stage prospects who want to be first to test this. If this sounds relevant to how your team works, feel free to reach out or drop me a message.
A note on trust
We’re aware that “AI in banking” sounds abstract to some and alarming to others. Our philosophy is clear: AI at amnis is designed to act within boundaries that humans define and approve. Every automated action has a human in the loop — whether that’s setting the parameters, reviewing the output, or approving the transaction.
The SMEs adopting AI fastest are also the ones thinking hardest about where human judgment is irreplaceable. We’re building for them.
