Connect with us

Accounting

Microsoft researchers teach LLMs to use spreadsheets well

Published

2 years ago

on

Large language models like ChatGPT have traditionally had trouble reading and interacting with spreadsheets, limiting their application in this realm, but recent research from Microsoft claims to have found an answer. 

The paper, SPREADSHEETLLM: Encoding Spreadsheets for Large Language Models, described the problems LLMs typically face with spreadsheets and proposed what it called the “SheetCompressor” framework to address them. 

The issue LLMs have with spreadsheets has to do with tokenage requirements. LLMs, generally, run on “tokens,” which are the basic units of data the model processes. Tokens are words, character sets, or combinations of words and punctuation that are used by large language models to decompose text into. LLMs operate by converting input text into a series of tokens, which the model then uses to understand and generate responses. 

The number of tokens determines the computational cost and capacity needed to handle the input, making token management crucial, especially for complex data like spreadsheets. For example, the phrase, “I heard a dog bark loudly at a cat” would be represented by eight tokens, one for each unique word. In order to preserve system resources, many LLMs have token limits, but even in a limitless environment, complex jobs are resource intensive, with significant computational effort that affects both performance and efficiency. 

Typically, each part of a spreadsheet — even blank cells or repeating cells or those with irrelevant information — costs tokens, meaning even a simple spreadsheet has a much higher token requirement than traditional text. Furthermore, LLMs often struggle with spreadsheet-specific features such as cell addresses and formats, complicating their ability to effectively parse and utilize spreadsheet data. These challenges have limited just how much generative AI models can be applied to reading and interacting with spreadsheets. Considering how many spreadsheets the profession tends to use, this consequently limits their application towards deep accounting work. 

What Microsoft researchers discovered, in short, is that the LLM does not need to burn tokens reading and processing the entire spreadsheet. Instead people can create a compressed version of the document to function as something like an index, with markers or “anchors” indicating especially important information like totals. Additional compression comes from grouping together similar types of data like date columns. So, in a sense, the LLM does not work through the spreadsheet itself but instead references it via a much more efficient index. 

Complex spreadsheets are further supported through a concept called “chain of spreadsheet,” which is similar to “chain of thought” prompting. The method unfolds in two stages. First, the model identifies the table that is relevant to the query and determines the precise boundaries of the relevant content. This step ensures that only pertinent data is considered in the subsequent analysis. Then, the query and the identified table section are re-input into the LLM. The model then processes this information to generate an accurate response to the query.

“Through the CoS, SPREADSHEETLLM effectively handles complex spreadsheets by breaking down the process into manageable parts, thus enabling precise and context-aware responses,” said the paper. 

Experiments with this method found that it significantly increased performance on larger spreadsheets where token limits are a particular challenge. The F1 score (which is used to measure the accuracy of an AI model) for massive spreadsheets was 75% higher than GPT-4 and 19% higher than TableSense-CNN, another spreadsheet methodology for AI; for large spreadsheets, the difference was 45% and 17% respectively; for medium spreadsheets it was 13% and 5%; and for small spreadsheets it was 8%. Overall, the results show that while the method gets more effective the larger the spreadsheet, it can still improve the efficiency of even small spreadsheets. 

“Through a novel encoding method, SHEETCOMPRESSOR, this framework effectively addresses the challenges posed by the size, diversity, and complexity inherent in spreadsheets,” the paper concluded. “It achieves a substantial reduction in token usage and computational costs, enabling practical applications on large datasets. The fine-tuning of various cutting-edge LLMs further enhances the performance of spreadsheet understanding. Moreover, Chain of Spreadsheet, the framework’s extension to spreadsheet downstream tasks illustrates its broad applicability and potential to transform spreadsheet data management and analysis, paving the way for more intelligent and efficient user interactions.”

Implications

Donny Shimamoto, founder and managing director of accounting tech-focused accounting firm IntrapriseTechKnowlogies said, by enabling LLMs to “understand” tabular spreadsheets, accountants will have increased ability to either summarize or analyze a set of data. More than that, however, he said this will likely allow even non-accountants to do the same, removing the accountant as the middle person. However while some accountants may see this as a threat, he said what this would mainly do is clear the majority of simple inquiries from their plates, letting them save their energy for more complex questions and deeper analysis.

“Implementing something like this will require good testing to ensure that the risk of hallucinations is minimized, especially if it is going to help provide non-accountants with information to support decision-making,” said Shimamoto.

David Wood, a Bringham Young University accounting professor who specializes in AI within the profession, raised a similar point, as it would allow those without significant technical knowledge to do the same kinds of tasks that, previously, could only be done by seasoned accounting experts. He raised the example of novices being able to use generative AI to make spreadsheets that only expert professionals could put together. However, while he thinks this could be possible soon, he said that, despite the Microsoft research, it hasn’t arrived just yet.

“However, there are at least three challenges holding back using GenAI with spreadsheets: the size and complexity of the spreadsheets, and the required accuracy for most uses of spreadsheets. This paper takes a large step in the right direction, but it doesn’t solve all the challenges and more work will still be needed in each of these three areas. It would be a mistake to assume that after reading this paper, we have fully figured out how to use spreadsheets and GenAI together. More work is still needed. … I think the path these researchers are taking is significant, but the research “hasn’t arrived yet” meaning that more work is needed. The accuracy rates are just not high enough…yet. Hopefully this paves the way for the next researcher to move it forward further.” he said in an email.

Related Topics:
Continue Reading

Accounting

Mandatory ESG Reporting Standards Demand Standardized Non-Financial Audit Trails

Published

4 days ago

on

September 5, 2026

By

Mandatory ESG Reporting Standards Demand Standardized Non-Financial Audit Trails

Corporate accounting departments face an expanded regulatory mandate as mandatory sustainability and Environmental, Social, and Governance (ESG) reporting frameworks take full effect internationally. Governed by the European Union’s Corporate Sustainability Reporting Directive (CSRD) and the International Sustainability Standards Board (ISSB) IFRS S1 and S2 standards, enterprise financial controllers are now legally required to track, verify, and report non-financial data with the same internal controls and auditability as traditional financial statements.

The expansion shifts ESG compliance

This regulatory expansion shifts ESG compliance from marketing departments to corporate accounting offices. Financial managers are now responsible for gathering, consolidating, and verifying carbon emissions metrics, supply chain labor conditions, water usage, and climate risk exposures across multi-tiered corporate structures. These non-financial metrics must be integrated into standardized general ledgers to withstand rigorous third-party audit assurance processes.

To comply with these rigorous reporting mandates, accounting software providers have added dedicated ESG modules designed to aggregate data from IoT sensors, utility platforms, and vendor management systems. Controllers are implementing internal control frameworks—modeled after traditional COSO frameworks—to ensure the completeness, accuracy, and consistency of sustainability disclosures, protecting organizations against greenwashing penalties and litigation risks.

The transition requires significant cross-functional collaboration between accounting teams, legal counsel, and operational directors. Accounting professionals are expanding their technical expertise beyond financial ledgers to master carbon accounting methodologies, lifecycle assessment standards, and non-financial data governance protocols, fundamentally expanding the role of the modern corporate accountant.

Why This Information Matters
Mandatory ESG disclosures require companies to treat environmental and social metrics as audited financial records. Executives, accountants, and board members must institute formal tracking and assurance processes to satisfy legal mandates, maintain investor confidence, and mitigate regulatory non-compliance risks.

Continue Reading

Accounting

SEC’s Semiannual Reporting Proposal Faces Investor Pushback: What CFOs Need to Know

Published

2 weeks ago

on

August 25, 2026

By

U.S. Securities and Exchange Commission (SEC)

A proposal from the U.S. Securities and Exchange Commission to potentially shift some public companies away from quarterly financial reporting toward a semiannual model is drawing significant pushback from investors, even as it continues moving through the regulatory process. The debate has direct implications for corporate finance teams, auditors, and the broader transparency of U.S. capital markets.

What the SEC Proposed

According to a summary published by accounting advisory firm Cohen & Co., the SEC issued a proposed rule on May 19, 2026, aimed at simplifying financial reporting requirements for many U.S. public companies. The proposal would potentially reduce the frequency of certain mandatory disclosures from quarterly to semiannual, a structural change that has not been made to core U.S. reporting requirements in decades.

The proposal follows an extended debate within U.S. policy circles, with proponents arguing that reduced reporting frequency could lower compliance costs and free up management time for longer-term strategic planning rather than quarter-to-quarter results management.

Why Investors Are Pushing Back

Comment letters submitted in response to the proposal have been extensive, and according to Cohen & Co.’s review of the public record, investors “appear to be largely opposed” to the shift, viewing frequent interim reporting as a core benefit of U.S. capital markets relative to other jurisdictions.

Accounting and law firms have taken a more measured position, generally urging any changes to remain aligned with the Financial Accounting Standards Board (FASB), whose existing disclosure requirements and guidance are built around a quarterly reporting cadence. A shift to semiannual reporting without corresponding changes to FASB guidance could create friction between SEC filing requirements and GAAP-based disclosure expectations.

Lessons From the U.K. Experience

The debate is not without precedent. The United Kingdom moved away from mandatory quarterly reporting for listed companies in 2014, returning to a semiannual disclosure requirement. According to Cohen & Co.’s analysis, that experience offers a cautionary data point: there was no measurable increase in capital expenditure or R&D investment following the change, while analyst coverage of affected companies declined as reliable interim information became less available — a particular risk for smaller and newly public companies that rely on analyst coverage to maintain investor visibility.

Practical Implications for Finance Teams

Beyond the debate over disclosure philosophy, the proposal carries practical complications. Many companies have debt covenants and credit agreements structured around quarterly financial delivery; a shift to semiannual reporting could require renegotiating those terms. Reduced reporting frequency would also extend the “window of market silence” between disclosures, a factor that governance and investor-relations teams would need to manage carefully to avoid information asymmetry.

Separately, and unrelated to the reporting-frequency debate, the SEC and FASB have continued finalizing more routine updates this year. New Accounting Standards Updates are taking effect for December 31, 2026, fiscal year-ends covering income tax disclosures, credit loss measurement, induced debt conversions, and stock compensation, according to Eide Bailly’s review of 2026 ASU activity. Additional guidance on paid-in-kind dividends and environmental credits is also on the near-term horizon.

What to Watch Next

The semiannual reporting proposal remains in the comment and review phase, and no final rule has been adopted as of this writing. Finance leaders should monitor the SEC’s regulatory agenda for further movement, while treating the current quarterly reporting requirement as the operative standard until any final rule is issued and an effective date is set.

Given the extent of investor opposition documented in the comment file, a full shift to mandatory semiannual reporting appears more likely to result in either a scaled-back compromise or continued study rather than swift adoption — though the SEC’s ultimate direction remains uncertain.

 

Continue Reading

Accounting

AI-Driven Automation and Continuous Accounting Frameworks

Published

4 weeks ago

on

August 15, 2026

By

The accounting profession is undergoing a fundamental structural transition as enterprise finance departments shift from periodic month-end closes toward automated continuous accounting models. By integrating specialized machine learning algorithms directly into enterprise resource planning (ERP) platforms, chief accounting officers are transforming financial reporting from a retrospective exercise into a real-time operational asset.

The Shift from Periodic Close to Continuous Financial Reporting
Traditional accounting workflows heavily relied on manual data reconciliation, spreadsheet calculations, and multi-week closing cycles at the end of each fiscal period. In contrast, continuous accounting frameworks utilize automated software agents to process, validate, and post transactional data in real time as business activities occur.

Automated bank reconciliation tools cross-reference incoming bank feeds, invoice records, and purchase orders automatically. By resolving transactional variances instantly throughout the month, corporate accounting teams eliminate the traditional workload spikes associated with quarterly and annual closes.

Machine Learning in Audit Trails and Anomaly Detection
Advanced natural language processing (NLP) and machine learning tools are redefining internal audit and financial control environments. Automated systems analyze 100% of general ledger entries, identifying anomalous transactions, duplicate payments, and unauthorized journal entries in real time.

Rather than relying on random statistical sampling, corporate internal auditors can focus their attention on high-risk flags automatically surfaced by algorithmic monitoring platforms. This continuous risk assessment strengthens internal controls over financial reporting (ICFR) and significantly reduces fraud risk.

Evolving Roles for Accounting Professionals
As routine data entry and manual reconciliation tasks become fully automated, the skill set required for accounting professionals is shifting toward data analysis, system design, and strategic business advisory.
– Systems Governance: Accountants are increasingly responsible for monitoring algorithmic accuracy and managing data integration pipelines.
– Business Partnership: Finance professionals leverage real-time financial dashboards to advise operational leaders on margin management and working capital allocation.
– Regulatory Compliance Management: Accounting teams utilize automated platforms to ensure compliance with dynamic tax codes and international accounting standards.

Core Implementation Recommendations
1. Deploy Automated Reconciliation Tools: Integrate continuous transaction processing modules into existing enterprise ERP architectures.
2. Establish Algorithmic Governance Controls: Implement strict internal testing protocols to ensure automated accounting rules comply with GAAP/IFRS standards.
3. Reskill Accounting Teams: Invest in training finance staff on data analytics, workflow automation, and predictive financial modeling.

Continue Reading

Trending