The automation of financial reports, such as balance sheets, profit and loss statements or account statements, is a practical example of AI in banks. Improvement from the receipt of the document, through the rough recording, to the analysis and archiving of the data.
Difference between balance sheet and annual financial statements
In connection with AI, many institutions primarily talk about the goal of automatic balance sheet analysis, but this description is often not entirely accurate. The balance sheet is part of the annual financial statements or financial report and therefore only part of the data that can be used to assess creditworthiness.
In addition to the balance sheet, which gives the balance sheet its name and compares the use of funds (assets) and the source of funds (liabilities), many other components of an annual financial statement provide data to refine the balance sheet.
- Profit and loss account
- Statement of accounts
- Certificate
- List of attachments
- Corporate Social Responsibility or CSR Report
This technically motivated understanding of a financial report should strongly characterize every AI system that is used to record, process and analyze this multidimensional and usually less structured data. This is the only way to ensure long-term explainability of the overall system, even with the increasing complexity of the AI application, e.g. the expansion from HGB to IFRS or US GAAP.
How much time is required for processing?
The evaluation of a survey illustrates the considerable effort involved in analyzing balance sheets in financial institutions in accordance with KWG 18. The following points are particularly interesting:
Entering the annual financial statements
- More than half of the respondents (52%) state that the manual entry of financial statements is predominantly carried out by analysts. This means that highly qualified specialists spend a significant amount of time on administrative tasks.
- Only 10% have access to specialized data collectors for this task, and 7% use service providers. The low use of external service providers or internal data collectors shows that the task is mainly handled internally and by qualified analysts.
- 31% of the participants did not provide any information, which could indicate that some institutions have unclear processes or a mixed distribution of tasks.
Checking the values in the balance sheet tool before analysis
- 38% confirm that analysts also check the values. The dual function of input and control by analysts can further slow down the process and limit their availability for other tasks.
- Only 3% state that the verification is carried out by data collectors, and similarly few (2%) outsource this work to service providers.
- It is striking that 22% state that no review takes place at all, which indicates possible risks in quality assurance.
- Here, too, a large proportion (32%) remains unanswered, which could indicate process gaps or uncertainties.
Duration of entering an average annual financial statement
- Only 2% of respondents state that the input takes less than 10 minutes. This is a small proportion, which indicates that in most cases the input is time-consuming.
- For 21%, the entry takes 10-20 minutes, and 27% state that it even takes 20-30 minutes. This shows that a significant proportion of employees spend more than 20 minutes per transaction.
- 10% of respondents say it takes even longer than 30 minutes to enter, which underlines the considerable amount of work involved.
- In addition, 40% did not provide a response, which may indicate uncertainties regarding the processes or lack of standard times for this task.
Duration of the input value check
- According to 18% respondents, it takes less than 5 minutes to check the input values. A further proportion of 18% need 5-10 minutes for this.
- 13% state that the check takes 10-15 minutes, while 4% state an even longer duration.
- Once again, almost half of the participants (47%) did not provide any information, which underlines the heterogeneity of the testing processes or uncertainties in the procedure.
These detailed figures show that the manual effort involved in both entering and checking balance sheets is considerable. The majority of the tasks are performed by analysts and the duration of the processes varies greatly, which indicates potential for optimization in terms of efficiency and standardization.
The future of financial statement analysis
A new approach to automating process steps in the analysis of annual financial statements is software with AI. AI-supported systems are able to analyze large amounts of financial data in real time and identify patterns and trends - including in balance sheet analysis. This speeds up the analysis process and significantly minimizes the risk of human error.
Automation of the annual financial statement analysis with Konfuzio
The Helm & Nagel GmbH, offers IDP software in the form of the Konfuzio product, which can also cover such complex technical challenges in document processing. With the help of the established AI platform, you are able to automate process steps in the analysis of annual financial statements and achieve new benefits as a result.
Konfuzio applies advanced technologies such as machine learning, optical character recognition and intelligent AI algorithms to analyze financial statements
- to scan,
- to understand and
- to interpret,
whereby relevant information is efficiently extracted from the annual financial statement documents using a combination of AI models. With Konfuzio, it is possible to provide this extracted data from the annual financial statement documents in a structured form and to accelerate and simplify the balance sheet analysis process. In addition to the structured form as a technical interface, so-called REST API, the extracted information can also be subsequently checked, versioned and reliably archived by balance sheet auditors with a user-friendly interface.
Konfuzio offers its customers these services
- Text recognition - Konfuzio can extract text from any company's balance sheet documents and convert it into machine-readable data.
- Classification - The AI is able to classify relevant documents for the annual financial statement analysis according to various categories, making it easier to organize and access information.
- Pattern recognition - Konfuzio recognizes patterns in financial statements that are important for analyzing and forecasting company performance.
- Ease of use - Konfuzio offers an intuitive user interface that enables even non-experts or accounting professionals to perform balance sheet analyses without additional training.
Advantages of an automated annual financial statement analysis
The use of Konfuzio in financial statement analysis offers numerous advantages. These include
- Time saving - Automated analysis speeds up the process considerably.
- Precision - AI minimizes the risk of human error and provides accurate results on the financial statements.
- Real-time analysis - Konfuzio enables real-time monitoring of financial data.
- Fast decision making - Investors and management can make informed decisions on the basis of current data.
Use cases for process automation with AI
The following use cases illustrate AI-supported financial statement analysis in practice in various industries:
Credit scoring in the banking sector
Credit scoring is part of day-to-day business for banks and credit institutions. By using advanced AI algorithms and machine learning techniques, banks can optimize credit scoring. By extracting historical transaction data, creditworthiness information, key figures and proof of income, the appropriate AI model supports the creation of risk profiles. This enables a faster and more accurate assessment of creditworthiness, reduces default risks and improves the efficiency of the credit approval process.
Balance sheet audit for audit firms
Audit institutions face the challenge of conducting a comprehensive and accurate annual audit to ensure the compliance and financial integrity of companies. This is another area where AI comes into play. AI systems can automate financial statement audit processes by using services such as text recognition, pattern recognition and data analysis techniques. These automated checks are performed faster and more accurately than manual ones, helping to identify potential irregularities or errors in the financial statements that a human might miss. This enables audit institutions to use their resources more efficiently while increasing the quality of the audit.
Business accounting for tax consultancies
according to HGB SKR 04 and SKR 03
Companies with HGB accounting generally follow the standards of the German Commercial Code (HGB) in their financial accounting. The SKR 04 and SKR 03 charts of accounts play an important role here:
| Principle | Chart of accounts | Differentiation |
|---|---|---|
| Process organization principle | SKR 03 | The accounts are organized according to the operational processes. The order of the accounts follows the sequence of business transactions in the company. |
| Closing principle | SKR 04 | The arrangement of the accounts is based on the structure of the annual financial statements, which comprise the balance sheet and the income statement. |
Due to these principles, the account numbers of SKR 03 and SKR 04 differ from each other. KI is able to standardize business accounting according to the chart of accounts.
What data can the AI provide for analysis?
The AI model extracts information from the annual financial statements completely automatically. If necessary, the extracted data can be adjusted manually:
Examples of extracted content, so-called labels, can be found in the table below with corresponding example values. Numerous other extractions are possible according to this scheme:
| LABEL | DESCRIPTION | TYPE | VALUE |
|---|---|---|---|
| Company name | The name of the bank for reconciliation. The company name is not normally shown on the balance sheet itself. The company name is usually listed at the top of the financial report or annual financial statements to identify the company. | Text | |
| Current income | 3. current income from a) b) c) Current income refers to regular income or earnings that are generated continuously over a certain period of time. | Number | |
| Total assets | This item comprises the total value of all assets that a company owns at a specific point in time. Assets are the resources that a company owns or controls and that serve to generate future economic benefits. Total assets is an important balance sheet item as it represents the value of the company's assets. | Number | 2.567.876,96 |
| Total liabilities and shareholders' equity | The "total liabilities" in a balance sheet is the total value of all liabilities and equity of a company at a specific point in time. Liabilities represent the company's financial obligations or debts to external parties. | Number | |
| Equity | 11 Equity (EK) is the difference between total assets (assets) and liabilities (liabilities, provisions), special items and deferred income (liabilities). | Number | |
Integration into existing IT structures and tools
With the REST API the AI modules from Konfuzio are modularly integrated into existing workflows, users are individually authorized and authenticated in accordance with industry standards, so-called single sign-on. The integration ranges from simple drag & drop and uploading by email to fully-fledged technical integration between the receiving system and the target system of the data for further processing.
Privacy and security
Data protection is a top priority at Konfuzio. All of our AI models process data exclusively in accordance with the GDPR and protect sensitive information throughout the entire extraction process. We also offer on-premise operation on the customer's servers in order to meet special requirements in terms of data protection. Data integrity implement.
Conclusion
The use of AI to automate the analysis of financial statements offers the financial sector an efficient and precise solution. Combined with other modern technologies such as those used by Konfuzio, it enables fast processing of financial data from any accounting system, minimizing errors and saving time. The diverse functions of the AI platform and its possible integration into a company's existing IT structures make Konfuzio a valuable tool for making well-founded decisions based on current data within the company.
By complying with data protection standards, Konfuzio guarantees the security of sensitive information. Overall, automating the analysis of annual financial statements offers a forward-looking opportunity for companies to optimize their analysis processes and strengthen their competitiveness.
EN 
DE 
ES 
FR