Connect with us

Accounting

Microsoft researchers teach LLMs to use spreadsheets well

Published

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.

Continue Reading

Accounting

XcelLabs launches to help accountants use AI

Published

on

Jody Padar, an author and speaker known as “The Radical CPA,” and Katie Tolin, a growth strategist for CPAs, together launched a training and technology platform called XcelLabs.

XcelLabs provides solutions to help accountants use artificial technology fluently and strategically. The Pennsylvania Institute of CPAs and CPA Crossings joined with Padar and Tolin as strategic partners and investors.

“To reinvent the profession, we must start by training the professional who can then transform their firms,” Padar said in a statement. “By equipping people with data and insights that help them see things differently, they can provide better advice to their clients and firm.”

Padar-Jody- new 2019

Jody Padar

The platform includes XcelLabs Academy, a series of educational online courses on the basics of AI, being a better advisor, leadership and practice management; Navi, a proprietary tool that uses AI to help accountants turn unstructured data like emails, phone calls and meetings into insights; and training and consulting services. These offerings are currently in beta testing.

“Accountants know they need to be more advisory, but not everyone can figure out how to do it,” Tolin said in a statement. “Couple that with the fact that AI will be doing a lot of the lower-level work accountants do today, and we need to create that next level advisor now. By showing accountants how to unlock patterns in their actions and turn client conversations into emotionally intelligent advice, we can create the accounting professional of the future.”

Tolin-Katie-CPA Growth Guides

Katie Tolin

“AI is transforming how CPAs work, and XcelLabs is focused on helping the profession evolve with it,” PICPA CEO Jennifer Cryder said in a statement. “At PICPA, we’re proud to support a mission that aligns so closely with ours: empowering firms to use AI not just for efficiency, but to drive growth, value and long-term relevance.”

Continue Reading

Accounting

Accounting is changing, and the world can’t wait until 2026

Published

on

The accountant the world urgently needs has evolved far beyond the traditional role we recognized just a few years ago. 

The transformation of the accounting profession is not merely an anticipated change; it is a pressing reality that is currently shaping business decisions, academic programs and the expected contributions of professionals. Yet, in many areas, accounting education stubbornly clings to outdated, overly technical models that fail to connect with the actual demands of the market. We must confront a critical question: If we continue to train accountants solely to file tax reports, are we truly equipping them for the challenges of today’s world? 

This shift in mindset extends beyond individual countries or educational systems; it is a global movement. The recent announcement of the CIMA/CGMA 2026 syllabus has made it unmistakably clear: merely knowing how to post journal entries is insufficient. Today’s accountants are required to interpret the landscape, anticipate risks and act with strategic awareness. Critical thinking, sustainable finance, technology and human behavior are not just supplementary topics; they are essential components in the education of any professional seeking to remain relevant. 

The CIMA/CGMA proposal for 2026 is not just a curriculum update; it is a powerful manifesto. This new program positions analytical thinking, strategic business partnering and technology application at the core of accounting education. It unequivocally highlights sustainability, aligning with IFRS S1 and S2, and expands the accountant’s responsibilities beyond mere numbers to encompass conscious leadership, environmental impact and corporate governance. 

The current changes in the accounting profession underscore an urgent shift in expectations from both educators and employers. Today, companies of all sizes and industries demand accountants who can do far more than interpret balance sheets. They expect professionals who grasp the deeper context behind the numbers, identify inconsistencies, anticipate potential issues before they escalate into losses, and act decisively as a bridge between data and decision making. 

To meet these expectations, a radical mindset shift is essential. There are firms still operating on autopilot, mindlessly repeating tasks with minimal critical analysis. Likewise, many academic programs continue to treat accounting as purely a technical discipline, disregarding the vital elements of reflection, strategy and behavioral insight. This outdated approach creates a significant mismatch. While the world forges ahead, parts of the accounting profession remain stuck in the past. 

The consequences of this shift are already becoming evident. The demand for compliance, transparency and sustainability now applies not only to large corporations but also to small and mid-sized businesses. Many of these organizations rely on professionals ill-equipped to drive the necessary changes, putting both business performance and the reputation of the profession at risk. 

The positive news is that accountants who are ready to thrive in this new era do not necessarily need additional degrees. What they truly need is a commitment to awareness, a dedication to continuous learning, and the courage to step beyond their comfort zones. The future of accounting is here, and it is firmly rooted in analytical, strategic and human-oriented perspectives. The 2026 curriculum is a clear indication of the changes underway. Those who fail to think critically and holistically will be left behind. 

In contrast, accountants who see the big picture, understand the ripple effects of their decisions, and actively contribute to the financial and ethical health of organizations will undeniably remain indispensable, anywhere in the world.

Continue Reading

Accounting

Republicans push Musk aside as Trump tax bill barrels forward

Published

on

Congressional Republicans are siding with Donald Trump in the messy divorce between the president and Elon Musk, an optimistic sign for eventual passage of a tax cut bill at the root of the two billionaires’ public feud.

Lawmakers are largely taking their cues from Trump and sticking by the $3 trillion bill at the center of the White House’s economic agenda. Musk, the biggest political donor of the 2024 cycle, has threatened to help primary anyone who votes for the legislation, but lawmakers are betting that staying in the president’s good graces is the safer path to political survival.

“The tax bill is not in jeopardy. We are going to deliver on that,” House Speaker Mike Johnson told reporters on Friday.

“I’ll tell you what — do not doubt, don’t second guess and do not challenge the President of the United States Donald Trump,” he added. “He is the leader of the party. He’s the most consequential political figure of our time.”

A fight between Trump and Musk exploded into public view this week. The sparring started with the tech titan calling the president’s tax bill a “disgusting abomination,” but quickly escalated to more personal attacks and Trump threatening to cancel all federal contracts and subsidies to Musk’s companies, such as Tesla Inc. and SpaceX which have benefitted from government ties.

Republicans on Capitol Hill, who had —  until recently — publicly embraced Musk, said they weren’t swayed by the billionaire’s criticism that the bill cost too much. Lawmakers have refuted official estimates of the package, saying that the tax cuts for households, small businesses and politically important groups — including hospitality and hourly workers — will generate enough economic growth to offset the price tag.

“I don’t tell my friend Elon, I don’t argue with him about how to build rockets, and I wish he wouldn’t argue with me about how to craft legislation and pass it,” Johnson told CNBC earlier Friday.

House Budget Committee Chair Jodey Arrington told reporters that House lawmakers are focused on working with the Senate as it revises the bill to make sure the legislation has the political support in both chambers to make it to Trump’s desk for his signature. 

“We move past the drama and we get the substance of what is needed to make the modest improvements that can be made,” he said.

House fiscal hawks said that they hadn’t changed their prior positions on the legislation based on Musk’s statements. They also said they agree with GOP leaders that there will be other chances to make further spending cuts outside the tax bill. 

Representative Tom McClintock, a fiscal conservative, said “the bill will pass because it has to pass,” adding that both Musk and Trump needed to calm down. “They both need to take a nap,” he said.

Even some of the House bill’s most vociferous critics appeared resigned to its passage. Kentucky Representative Thomas Massie, who voted against the House version, predicted that despite Musk’s objections, the Senate will make only small changes.

“The speaker is right about one thing. This barely passed the House. If they muck with it too much in the Senate, it may not pass the House again,” he said.

Trump is pressuring lawmakers to move at breakneck speed to pass the tax-cut bill, demanding they vote on the bill before the July 4 holiday. The president has been quick to blast critics of the bill — including calling Senator Rand Paul “crazy” for objecting to the inclusion of a debt ceiling increase in the package.

As the legislation worked its way through the House last month, Trump took to social media to criticize holdouts and invited undecided members to the White House to compel them to support the package. It passed by one vote.

Senate Majority Leader John Thune — who is planning to unveil his chamber’s version of the bill as soon as next week — said his timeline is unmoved by Musk. 

“We are already pretty far down the trail,” he said.

Continue Reading

Trending