Maaliya app is now live. Get access now and book a demo here.

AI & Automated Processing Policy

Last updated 28 Aug 2026

Maaliya Accounting and Bookkeeping DWC-LLC ("Maaliya", "we", "us" or "our") uses artificial intelligence, machine learning, optical character recognition ("OCR"), rules-based automation and other automated technologies to make accounting and finance workflows faster and more efficient.

This AI & Automated Processing Policy ("Policy") explains how those technologies are used, when human or customer review is expected, how third-party AI providers may be involved, and the limitations customers should understand when relying on automated outputs.

Maaliya is designed to assist businesses and finance teams — not to remove appropriate human judgement from accounting, tax or financial decisions.

1. Scope

This Policy applies to automated and AI-assisted functionality available through:

  • maaliya.ae;
  • app.maaliya.ae;
  • Maaliya mobile applications;
  • Maaliya accounting and bookkeeping services; and
  • related Maaliya products and integrations.

Not every feature described in this Policy is available to every customer, subscription or service plan.

This Policy should be read alongside Maaliya's Terms of Service, Privacy & Cookie Policy and Data Processing & Subprocessors Policy.

If this Policy conflicts with the Terms of Service or a separately executed agreement, the applicable contractual agreement will prevail.

2. How Maaliya Uses Automation and AI

Maaliya combines different technologies rather than relying on a single artificial intelligence system.

Depending on the feature being used, processing may involve:

  • Maaliya-developed systems;
  • rules-based automation;
  • OCR and document-processing technology;
  • reconciliation algorithms;
  • machine-learning or AI models;
  • confidence scoring;
  • third-party AI services; and
  • human review where included within the customer's service.

The majority of Maaliya's OCR and reconciliation processing is performed using Maaliya's own systems or specialised tools rather than general-purpose AI models.

General-purpose AI may be used in limited or edge cases where additional interpretation or processing is useful.

3. Document Processing and OCR

Maaliya may use OCR and other document-processing technologies to extract information from documents including:

  • invoices;
  • bills;
  • receipts;
  • PDFs;
  • photographs and scanned documents;
  • bank statements; and
  • other financial or accounting records.

Depending on the document, Maaliya may attempt to identify or extract information such as:

  • supplier or customer details;
  • invoice numbers;
  • invoice and due dates;
  • transaction dates;
  • amounts;
  • VAT amounts;
  • Tax Registration Numbers (TRNs);
  • line items;
  • bank transactions;
  • descriptions; and
  • other accounting information.

OCR is an automated interpretation of the document provided to Maaliya. It does not replace the original document.

Customers are responsible for ensuring that documents submitted to Maaliya are complete, accurate and sufficiently readable.

4. Accounting Categorisation

Maaliya may automatically analyse transactions and documents and assign or suggest accounting information.

This can include:

  • accounting categories;
  • chart-of-account codes;
  • suppliers and customers;
  • transaction descriptions;
  • tax treatments;
  • bookkeeping classifications; and
  • other accounting fields.

Maaliya may use previous transactions, rules, document information, customer configurations and automated models to generate these results.

Customers can review and, where functionality permits, change automated classifications.

5. VAT and Tax Suggestions

Maaliya may automatically analyse transactions and documents to determine or suggest VAT and other tax treatments.

For example, Maaliya may identify a transaction as:

  • standard-rated;
  • zero-rated;
  • exempt;
  • outside the scope of VAT;
  • subject to reverse charge; or
  • requiring review.

Maaliya may also provide Corporate Tax-related classifications, calculations, suggestions or other tax functionality.

These outputs are intended to assist accounting and compliance workflows. Automated tax classifications and suggestions do not constitute legal or tax advice and should not be treated as a substitute for professional judgement where professional advice is appropriate.

Customers remain responsible for reviewing the accuracy of relevant tax information and ensuring that filings and submissions made to tax authorities are complete and correct.

Maaliya will not submit a VAT or Corporate Tax filing to the UAE Federal Tax Authority on behalf of a customer without the customer's appropriate authorisation.

6. Reconciliation

Maaliya may automatically analyse financial information to identify potential relationships between transactions.

This may include matching:

  • bank transactions with invoices;
  • bank transactions with bills;
  • payments with customers or suppliers;
  • transactions with accounting records; and
  • other financial records.

Suggested matches may be based on information such as amounts, dates, references, counterparties, historical activity and other relevant data.

Confidence levels and review states may be used to distinguish between stronger matches and transactions requiring further attention.

7. Automated Actions and Customer Rules

Some Maaliya functionality can perform actions automatically without requiring individual approval for every transaction.

However, automation of this kind is controlled by customer instructions.

For example, a customer may create or enable a rule allowing Maaliya to automatically process, categorise, reconcile or post transactions relating to a specific vendor or contact.

Maaliya will only bypass individual approval in these workflows where the customer has expressly enabled the relevant automation for that vendor, contact or applicable rule.

Customers can manage these permissions through the available rules and automation settings.

By enabling an automated rule, the customer instructs Maaliya to perform the relevant actions according to that rule until it is changed or disabled.

Customers remain responsible for reviewing their automation rules and ensuring they remain appropriate.

8. Identifying Automated and User-Provided Decisions

Maaliya's interface is designed to help users understand whether certain accounting information was generated automatically or entered by the user.

Where supported by the relevant interface:

Green outlines indicate information or decisions generated by Maaliya's automated systems.

Yellow outlines indicate information entered or selected by the user.

This distinction is intended to make automated processing visible and help users identify information they may wish to review.

Users can change automated outputs where the relevant functionality permits.

Interface designs and visual indicators may evolve as Maaliya's products develop, and equivalent indicators may be used in future versions of the Services.

9. Confidence Scores and "Needs Review"

Maaliya may assign confidence scores or other indicators to automated results.

Where Maaliya determines that an automated result requires additional attention, it may flag the item as "Needs Review" or use another equivalent review indicator.

Confidence scores are intended to assist users in prioritising review. They are not guarantees of accuracy.

An output may still be incorrect even where Maaliya assigns it a high confidence score.

Users should therefore apply appropriate judgement before relying on automated results.

10. Approval and Posting

In most Maaliya workflows, users will have an opportunity to review relevant accounting information before it is approved or posted to an accounting system such as Xero or Zoho.

Where a customer has expressly enabled an automation rule, certain actions may occur automatically as described in this Policy.

Customers are responsible for understanding which automation rules they have enabled and for reviewing those rules periodically.

Where a customer approves an automated suggestion after having an opportunity to review it, the customer's approval constitutes its instruction to proceed with that information.

The allocation of responsibility and liability relating to use of the Services remains subject to Maaliya's Terms of Service and any applicable customer agreement.

11. Self-Service Customers

For customers using Maaliya on a self-service basis, the customer is the human review layer.

Automated outputs are provided to assist the customer's accounting and finance processes, but the customer remains responsible for reviewing outputs before relying on them for material accounting, tax, financial or business purposes.

This includes reviewing relevant outputs against underlying records such as:

  • invoices;
  • receipts;
  • bank statements;
  • contracts;
  • accounting records; and
  • other supporting documents.

Customers should correct automated results where they identify an error or inconsistency.

12. Done-For-You Services and Human Review

Customers using Maaliya's Done-For-You ("DFY") accounting services receive an additional human review layer as part of that service.

For DFY customers, Maaliya account managers may review, verify and correct automated accounting outputs as part of providing the contracted accounting or bookkeeping service.

The combination of automated processing and professional human review is a feature of the DFY service.

This does not mean that every automated output across every Maaliya product receives human review.

Customers using self-service products should not assume that Maaliya personnel have reviewed an automated result unless the relevant service expressly includes such review.

13. AI Assistant and Financial Insights

Maaliya may provide conversational or AI-assisted functionality that can analyse information within a customer's account.

Depending on the features available, users may be able to ask questions concerning their financial information, including matters such as:

  • unpaid or overdue invoices;
  • receivables;
  • spending;
  • transactions;
  • cash flow;
  • accounting records;
  • VAT information; and
  • other financial summaries or insights.

These systems may interpret both the user's question and underlying financial information using automated technologies.

Conversational AI can misunderstand a question, context or underlying data.

Important answers should therefore be verified against the customer's underlying accounting records and supporting documentation before being relied upon.

14. Third-Party AI Providers

Maaliya currently uses OpenAI as a third-party general-purpose AI provider for certain limited or edge-case processing.

Maaliya may add, remove or replace AI providers as its Services and technology evolve and is not required to provide advance notice of every provider change unless required by applicable law or an applicable customer agreement.

Where third-party AI is used, Maaliya seeks to limit the information transmitted to what is reasonably necessary for the relevant function.

Third-party AI providers process information according to Maaliya's applicable commercial arrangements with them and the provider terms and data-processing arrangements applicable to the service being used.

Further information regarding third-party processing is available in Maaliya's Data Processing & Subprocessors Policy.

15. Use of Data to Improve Maaliya

Maaliya may use aggregated or anonymised information derived from use of the Services to:

  • evaluate system performance;
  • improve automation;
  • improve categorisation and reconciliation;
  • develop features;
  • analyse usage patterns;
  • improve internal models and systems; and
  • improve the overall Maaliya service.

Maaliya does not use identifiable Customer Data to train general Maaliya AI models.

Information used for general model improvement or training will be aggregated, anonymised or otherwise processed so that it does not reasonably identify the customer or an individual.

Where Customer Data is processed by a third-party AI provider, the provider's handling of that information is governed by the applicable commercial, contractual and data-processing arrangements.

16. Automated Decision-Making

Maaliya uses automated systems to make and recommend accounting, categorisation, reconciliation and tax-related decisions within business workflows.

These are operational accounting decisions and recommendations.

Maaliya does not use automated processing to make decisions about individuals that produce legal effects or similarly significant effects of the kind regulated as significant automated decision-making under applicable data-protection laws.

Where a workflow could materially affect a customer's accounting or tax position, users should exercise appropriate review and judgement before relying on the output.

17. Limitations of AI and Automated Systems

Artificial intelligence, OCR and automated systems are probabilistic and may produce incorrect, incomplete, outdated or unexpected results.

A result generated by Maaliya should not be assumed to be correct solely because:

  • it was generated automatically;
  • the system assigned a high confidence score;
  • similar transactions were previously processed in the same way; or
  • the system did not flag the item for review.

Automated systems also depend heavily on the quality and completeness of the information provided to them.

Maaliya does not guarantee that automated functionality will identify every accounting issue, tax issue, inconsistency or problem contained within Customer Data.

18. Customer Responsibility

Customers remain responsible for the accuracy and completeness of the information they provide to Maaliya.

Customers using automated functionality are also responsible for:

  • reviewing material automated outputs;
  • checking outputs against appropriate supporting documentation;
  • correcting identified errors;
  • reviewing transactions marked as requiring attention;
  • configuring automation rules appropriately;
  • monitoring rules that permit automatic actions;
  • maintaining accurate accounting records; and
  • obtaining professional advice where appropriate.

Maaliya's automated features should not be used as a substitute for a qualified accountant, tax adviser, lawyer or other professional where the circumstances require professional judgement or advice.

The fact that Maaliya provides accounting automation does not remove the customer's responsibility to ensure that its accounting, tax and regulatory obligations are properly met.

19. Requesting Human Assistance

If a user believes an automated result is incorrect, the user may correct the result directly where functionality permits or contact Maaliya for support.

Maaliya may manually inspect an automated result where reasonably necessary for:

  • customer support;
  • troubleshooting;
  • quality control;
  • investigating unexpected behaviour; or
  • providing a contracted managed service.

Any such access remains subject to applicable customer agreements, access controls and Maaliya's data-processing practices.

Human review is not guaranteed for self-service transactions unless expressly included in the applicable service.

20. Changes to AI Features

AI and automated technologies evolve quickly.

Maaliya may modify:

  • AI models;
  • OCR systems;
  • reconciliation technology;
  • confidence methodologies;
  • automated workflows;
  • third-party AI providers;
  • review processes; and
  • the functionality available within different plans.

We may update this Policy where material changes affect how automated processing is used or how Customer Data is handled.

The date at the top of this Policy indicates when it was last updated.

21. Contact Us

Questions about Maaliya's use of artificial intelligence or automated processing can be directed to:

Maaliya Accounting and Bookkeeping DWC-LLC, Trade Licence No. TL 14812, Floor 3, Tower A3, Dubai South Business Park, Dubai, United Arab Emirates.

Email: support@maaliya.ae

Books that close themselves.
Built for the UAE.
Book a demo