Summary

• Talky's €1.5 million pre-seed round closed in September 2026, with participation from HelloWorld, Silence VC, Motive Force, and Aera Capital.
• The software uses accounting history and staff expertise to process documents and resolve exceptions within existing financial systems.
• Talky reports processing more than 1.8 million documents covering over €2 billion in customer expenditure.

Madrid-based Talky has disclosed a €1.5 million pre-seed investment to develop AI software that automates accounting work using an organisation’s financial history and the knowledge of its employees.

The round closed in September 2026, although the financing became public at the end of October. Investors included HelloWorld, Silence VC, Motive Force, and Aera Capital, alongside individual technology investors. The funding will support product development, technical infrastructure, recruitment, and the company’s international expansion.

Founded in 2025 by chief executive Jorge Beamonte and chief operating officer Maximilian Letón, Talky is developing software intended to handle routine financial transactions while learning how individual businesses make accounting decisions that are not captured in formal rules.

The company reports processing more than 1.8 million documents associated with over €2 billion in customer expenditure. It says that between 86% and 93% of the documents handled by its platform are processed without human intervention, although those figures describe its own operating results rather than an independently established industry benchmark.

Learning the rules behind accounting decisions

Many established accounting systems can extract information from invoices, apply predetermined classifications, and create entries in financial records. Difficulties arise when the information does not match an existing rule or when a decision depends on knowledge held by employees rather than stored in the system.

A supplier might use different descriptions for the same type of expenditure, for example, while a contract could require a particular tax treatment or allocation across several accounts. Finance employees often resolve these cases by examining previous transactions, checking supporting documents, or speaking to colleagues familiar with the arrangement.

Talky’s software attempts to make that accumulated knowledge available during subsequent processing. It examines historical accounting records, supplier information, contracts, and internal policies to identify how comparable transactions were treated in the past.

When a new document arrives, the platform can use those records to propose accounting entries and determine whether additional information is needed. Rather than depending entirely on fixed rules, it is designed to apply context drawn from the company’s existing financial operations.

This approach extends beyond document extraction. Recognising an invoice number and payment amount establishes what a document contains, whereas determining the appropriate account, approval, and tax treatment requires an understanding of how the transaction fits into a business’s procedures.

Making those decisions consistently is important for organisations processing large volumes of financial documents. Incorrect classifications can create reconciliation problems, complicate reporting, or require additional work during internal and external reviews.

Exceptions return to finance employees

When the software encounters a transaction it cannot confidently resolve, it can identify the appropriate employee and ask for a decision. The resulting response becomes part of the information available when similar transactions arise.

That allows the platform to build a record of company-specific accounting practices through routine use, rather than requiring every procedure to be written down before automation begins.

Talky also associates decisions with supporting information so that financial teams can review the basis for an accounting entry. This traceability is relevant where businesses must establish who approved a transaction, which evidence was considered, and whether the treatment followed applicable procedures.

Its reported automation rate suggests that a substantial proportion of documents can be processed without direct employee involvement in the company’s existing deployments. However, the rate does not establish that every entry is correct or that the same performance will apply to businesses with different financial systems, document quality, and accounting policies.

As a company introduces the software, historical records may contain inconsistent classifications or procedures that are no longer appropriate. Using previous transactions to inform new decisions therefore requires controls that account for errors, policy changes, and exceptional circumstances.

Employees remain responsible for the quality of the organisation’s accounting processes, even where software performs much of the initial processing. Automated entries still need to comply with internal controls and any relevant financial reporting requirements.

Connecting automation to existing financial systems

Talky is designed to work alongside enterprise resource planning systems rather than replace the applications holding a company’s financial records. Its reported integrations include established platforms such as SAP and Microsoft Dynamics 365, allowing the software to operate within existing finance environments.

Connecting directly to those systems can reduce repeated data entry, but implementation depends on the quality of available records and the permissions granted to the automation platform. Businesses must decide which transactions can be posted automatically, which require approval, and how employees can investigate or reverse an incorrect entry.

Accounts payable, accounts receivable, and bank reconciliation provide potential applications because they involve recurring documents and transactions that can be compared against existing records.

Bank reconciliation, for example, requires businesses to match payments with invoices and other entries. The work becomes more demanding when one payment covers several documents, amounts differ because of adjustments, or identifying information is incomplete.

Software that understands previous decisions and supplier relationships may help resolve these cases, although unusual transactions still require appropriate accounting judgment.

Large enterprises also need to maintain consistent financial controls across departments and subsidiaries. The commercial value of the platform will depend partly on whether company-specific knowledge can be applied reliably without allowing informal exceptions to become uncontrolled accounting practices.

Funding supports expansion beyond initial customers

Talky operates in a growing European market for AI-based financial software, with companies developing tools for invoice processing, reconciliation, tax preparation, and other accounting activities.

Its approach concentrates on the institutional knowledge surrounding transactions, which may be harder to capture than the information printed on a document. That creates a potential distinction from systems focused primarily on recognising fields or applying predetermined classifications.

However, enterprise customers must weigh the benefits of automated processing against the cost of integration, oversight, and maintaining accurate decision records. Improvements in document-processing speed do not necessarily translate into equivalent financial savings if additional review is required elsewhere.

With a team of nine employees at the time of the financing disclosure, Talky plans to increase product and engineering capacity while strengthening relationships with existing customers. Its longer-term expansion will depend on maintaining performance across organisations with different accounting requirements.

The September investment provides additional resources for that work. As the company expands internationally, the consistency and auditability of its accounting decisions will be central to establishing whether the platform can operate across larger and more complex finance departments.