Xero Banned AI From Its API: What It Means for the Future of Accounting Software
In March 2026, Xero prohibited developers from using API data to train AI models and introduced per-connection fees. Here is why this matters for every UK business using Xero.
Quick Answer
Xero banned developers from using API data to train AI models in March 2026, killed its revenue-share model, and introduced per-connection fees. Three connected apps raised prices immediately. This signals that Xero sees AI-native accounting tools as a competitive threat and is choosing to restrict innovation rather than enable it.
What Xero Did in March 2026
On 2 March 2026, Xero updated its API terms to explicitly prohibit any use of data accessed through its API for training artificial intelligence or machine learning models. At the same time, Xero retired its longstanding revenue-share model for app partners and replaced it with a new fee structure: per-connection charges plus data egress fees.
The changes took effect immediately. Within the same week, three Xero-connected apps raised their prices to cover the new costs.
For the estimated four million businesses worldwide using Xero, and the thousands of developers building on its platform, this was not a minor policy update. It was a strategic decision that reveals exactly how Xero views the future of accounting software, and its own position in it.
Why Xero Made This Decision
The timing was not accidental. Throughout 2025 and early 2026, a wave of AI-powered accounting tools emerged that could connect to Xero via its API, pull a business's transaction data, and use it to deliver capabilities that Xero itself does not offer: natural language conversations with your books, proactive tax advice, automated receipt processing, and intelligent expense categorisation.
These tools were, in effect, using Xero as a data layer while delivering the value on top. The AI became the interface. Xero became the database.
From Xero's perspective, this is an existential threat. If the value shifts from the accounting platform to the AI layer sitting on top of it, Xero's pricing power disappears. Why would a business pay £37-65 a month for Xero when the AI tool does everything they need and Xero is just a backend they never open?
The API ban is Xero's answer: make it legally and financially difficult for AI-native competitors to use Xero data. If you cannot train your AI on Xero data, and if connecting to Xero costs you per-connection fees, the economics of building on top of Xero become much harder.
What the Industry Expert Predicted
One widely cited industry analysis of the changes put it bluntly: AI-native tools will simply build their own general ledgers and stop treating Xero as the system of record entirely.
This prediction cuts to the heart of the issue. Xero's moat has always been the ledger, the canonical record of a business's financial position. Every bank feed, every invoice, every receipt flows into the Xero ledger, and that gravity keeps businesses locked in.
But if AI-native tools build their own ledger, that lock-in disappears. The business connects their bank directly (via Plaid or Open Banking), forwards receipts to the AI, and lets the AI maintain the books. Xero is no longer needed.
This is not theoretical. It is already happening.
The Data Ownership Question
The most uncomfortable aspect of Xero's decision is what it implies about data ownership. When a business connects to Xero and enters its transactions, invoices, and receipts, whose data is it?
The business would reasonably say: "It is my financial data. I created it. I should be able to use it however I want, including letting an AI tool analyse it."
Xero's API terms now say: "You can access this data through our API, but you cannot use it to train AI models."
This creates a strange situation. A business can export its own data as a CSV and do whatever it wants with it. But if a tool accesses the same data through the API, it is restricted. The limitation is not about data ownership, it is about control. Xero controls the API, and they are using that control to limit what competitors can do with business data that does not belong to Xero.
For UK businesses evaluating their accounting software, this raises an important question: do you want your financial data held by a platform that actively restricts what you can do with it?
The Per-Connection Fee Impact
Beyond the AI training ban, the new per-connection pricing model has immediate financial consequences for the Xero ecosystem.
Under the old model, app partners shared revenue with Xero when their customers were also Xero subscribers. This aligned incentives: more apps meant more value for Xero users, which meant more Xero subscriptions.
The new model charges app developers per connection and per data request. This means:
- Apps that sync frequently cost more. A bank feed that checks for new transactions every hour now has a marginal cost per request. Developers either absorb this cost or pass it on to users.
- Small developers are priced out. The revenue-share model had no upfront costs. Per-connection fees require developers to pay Xero before they earn anything from their users.
- App prices go up. Three apps raised prices in the first week. More will follow as the economics settle.
The net effect is that the Xero ecosystem becomes more expensive and less innovative. Fewer developers will build on Xero, and those that remain will charge more. Xero users pay the cost.
What This Means for UK Businesses
If you are a UK limited company director or small business owner using Xero, here is the practical impact:
Your software costs will keep rising
Xero raised UK prices in September 2025. The Growing plan went from £33 to £37 per month. The Comprehensive plan (required for multi-currency) went to £50 per month. With the new API fees being passed through by app developers, the total cost of the Xero ecosystem, the core subscription plus connected apps, will continue to increase.
Innovation on Xero will slow down
The best new accounting tools are AI-powered. By banning AI training on API data and charging per-connection fees, Xero has made it harder and more expensive for these tools to exist. The apps that remain on the Xero marketplace will be the ones that can afford the fees, which typically means larger, established companies, not the innovative startups building the most useful features.
Your data is less portable than you think
While you can always export your data from Xero as a CSV, the API restrictions mean that seamless, automated migration to AI-native tools is harder. Xero is increasing switching costs, which is the classic sign of a platform defending a weakening competitive position.
The alternative is emerging
AI-native accounting tools that do not depend on Xero are being built right now. These tools connect directly to your bank (via Plaid or Open Banking), process receipts independently, and maintain their own ledger. They do not need Xero's API because they have replaced the functionality that Xero provides.
How AI-Native Tools Work Without Xero
The architecture of an AI-native accounting tool looks fundamentally different from a tool built on top of Xero:
Bank connections
Instead of relying on Xero's bank feed (which Xero can restrict or charge for), AI-native tools connect to banks directly through Open Banking (regulated by the FCA in the UK) or services like Plaid. The business authorises read-only access to their bank account, and transactions flow in automatically. No intermediary. No per-connection fees.
Receipt and invoice processing
Instead of uploading documents to Xero's receipt capture (or a connected app that accesses Xero), AI-native tools accept documents through multiple channels: email forwarding, photo upload, WhatsApp messages, or direct conversation. The AI extracts the data, categorises it, and matches it to bank transactions without any manual intervention.
The ledger
This is the key difference. AI-native tools maintain their own general ledger, built from the ground up for AI interaction. Every transaction carries rich metadata: the AI's categorisation confidence, the tax implications, the matched receipt, the relevant HMRC rule. This contextual data is what enables intelligent reasoning about your financial position, and it is something that cannot be retrofitted onto a legacy ledger.
Tax knowledge
Instead of relying on generic categorisation, AI-native tools are built with UK tax law embedded in the system. They know the current corporation tax rates, the optimal director salary, the rules for allowable expenses, and the deadlines for every filing obligation. This knowledge is not a plugin or an add-on. It is core to how the AI operates.
The Bigger Picture
Xero's API decision is a symptom of a larger shift in the accounting software market. The incumbents, Xero, QuickBooks, Sage, and FreeAgent, were all built in a pre-AI world. Their products assume human operators, their pricing assumes manual work, and their business models assume that the ledger is the most valuable asset.
AI changes all of these assumptions:
- Human operators become optional. AI can categorise, reconcile, and report without human intervention.
- Manual work disappears. The tasks that justify £50 per month subscriptions are now automated.
- The ledger is commoditised. A general ledger is a database. The value is in what you do with the data, not in storing it.
Xero's response is to restrict access and raise prices. The alternative response is to embrace AI and build something genuinely new.
Frequently Asked Questions
Why did Xero ban AI training on API data?
Xero banned the use of API data for AI and machine learning training in March 2026 because AI-native accounting tools were using Xero as a data source while delivering AI-powered features that compete with Xero's own product. By restricting AI training, Xero limits the ability of competitors to build intelligent tools on top of Xero data.
Can I still use AI tools with my Xero data?
You can still export your data from Xero as a CSV and use it with any tool you want. The restriction applies specifically to accessing data through the Xero API for AI training purposes. However, the new per-connection fees make it more expensive for AI tools to maintain real-time connections to Xero, which may reduce the availability or increase the cost of AI-powered Xero integrations.
Does this affect my Xero subscription price?
The API changes do not directly affect your Xero subscription price. However, connected apps that rely on the Xero API may raise their prices to cover the new per-connection and data egress fees. Three apps raised prices within the first week of the changes. Your total cost of using Xero plus its ecosystem of apps is likely to increase.
What is the alternative to Xero for UK businesses?
AI-native accounting tools like AccountsOS connect directly to your bank via Open Banking or Plaid, process receipts automatically, and maintain their own ledger with UK tax law built in. These tools do not depend on Xero's API and are typically cheaper (free during early access, then around £9-19 per month compared to Xero's £37-65 per month).
Can I migrate my data from Xero to an AI-native tool?
Yes. Most AI-native tools offer migration paths from Xero. You export your chart of accounts, transaction history, and contacts as CSV files from Xero, then upload them to the new platform. The AI maps your existing categories automatically. A typical migration takes under 30 minutes.
Who owns my financial data in Xero?
You own your financial data. Xero's terms of service confirm this. However, Xero controls the API through which third-party tools access your data, and they can set restrictions on how that API is used. The distinction is between owning the data and controlling access to the data, and Xero's recent changes are about tightening control over access.
Is Xero still a good choice for UK small businesses?
Xero remains a capable accounting platform with a large ecosystem of connected apps and strong accountant support. However, its pricing has increased significantly (the Growing plan is now £37 per month, up from £26 two years ago), its AI features are limited compared to AI-native alternatives, and its recent API restrictions suggest a defensive strategy rather than an innovative one. Whether it is the right choice depends on your specific needs, your budget, and how important AI-powered automation is to your workflow.
Will more accounting platforms restrict AI access?
It is possible. QuickBooks (Intuit) has its own AI strategy with Intuit Assist and may take similar steps to protect its data ecosystem. Sage has been less aggressive on AI. FreeAgent, owned by NatWest, has limited API access already. The trend in the industry is toward more restrictions on data access, which strengthens the case for using tools that connect to your financial data directly rather than through intermediary platforms.
Entrepreneur and technologist building AI-powered tools for UK small businesses.
Let AI handle your accounting
Stop worrying about deadlines and compliance. AccountsOS automates your bookkeeping so you can focus on growing your business.
Get Started Free