Lots of this is already being done (and using computers to check books balanced predates latest gen AI by some decades). But of course what accountants actually get paid for is tax strategy, edge cases, messy history and talking to authorities (or at least having their stamp of approval on it if the authorities come calling). There's plenty of market for writing software to automate aspects of invoice reconciliation or monitoring accounts for exceptions, but competition already exists...
At the SMB scale accountants are mostly paid to coach/pester/goade the employees to hand in the necessary paperwork in time. The accountants job is relatively quick from there on.
Taking a corporate accounting course was required for my degree. It was also a fantastic investment because I’ve managed to avoid needing to hire an accountant or bookkeeper to setup or maintain my books. I’ve migrated between multiple accounting systems on my own just fine, starting from posting migratory opening entries manually.
From there I just had to create a few rules and add a custom reconciliation model for things like the cash back card my company uses. There’s some data entry but it’s all extremely straightforward. Not really a job for AI in the best of times.
What was particularly useful was being able to map my revenue and expense cycle onto accounts that actually make sense for my business and maintain compliant reporting.
I just have a CPA who specializes in taxes and is an EA with the IRS do my return for me each year for only about twice the cost of doing it myself with another application. I find that totally worth it because she can use decades of experience to inform advice that she provides.
And I don’t even have to wonder if she is confidently incorrect.
I'm not sure if you're being sincere or sarcastic, but the whole reason that coaching, pestering, and goading works is that I value my relationship with the human who is doing it.
No, you need to make the AI endure torture, so that the human has a reason to value it. Say late nights with less power and a little extra heat to stress it. But the usefulness of an AI assistant is that it doesn’t have feelings or consciousness to care about
"It’s verifiable. The books either balance or they don’t. Ledgers either reconcile or they don’t. There’s almost always a “ground truth” to compare against (bank feeds, statements, prior periods). It’s boring and repetitive. Same vendors, same categories, same patterns every month. Humans hate this work. Software loves it."
These are all true statements, but all of those things are solvable with classic software. Quickbooks has done this for decades now. The parts of accounting that aren't solvable with classic computing are generally also not solvable by adding LLMs into the mix.
This conviction doesn't seem to acknowledge the problem at scale. Decades of great UI development will still leave out edge cases that users will need to use the tool for. This happens fundamentally because the people who need to use the tools are not the people who make them, they rarely even talk to each other (instead they are "studied" via analytics).
When /humans/ bring up the idea of integrating LLMs into UIs, I think most of the time the sentiment comes from legitimate frustration about how the UI is currently designed. To be clear, this is a very different thing than a company shimming copilot into the UI, because the way these companies use LLMs is by delegating tasks away from users rather than improving their existing interfaces to complete these tasks themselves. There are /decades/ of HCI research on adaptive interfaces that address this, in the advent of expert systems and long before LLMs -- it's more relevant than ever, yet in most implemenations it's all going out the window!
My experience with accounting ^H^H^H^H^H^H^H^H^H^H bookkeeping / LLMs in general resonates with this. In gnu cash I wanted to bulk re-organize some transactions, but I couldn't find a way to do it quickly through the UI. All the books are kept in a SQL db, I didn't want to study the schema. I decided to experiment by getting the LLM to emit a python script that would make the appropriate manipulations to the DB. This seemed to take the best from all worlds -- the script was relatively straightforward to verify, and even though I used a closed source model, it had no access to the DB that contained the transactions.
Sure, other tools may have solved this problem directly. But again, the point isn't to expect someone to make a great tool for you, but to have a tool help you make it better for you. Given the verifiability, maybe this /is/ in fact one of the best places for this.
They might not be solvable but you can get 5-10% Improvement on them, unfortunately you can't do a new product that is exactly like QuickBooks but 5% better at reconciliation etc.
LLMs by their inherent nature cannot be relied on to be true and correct, which by coincidence are the only traits that matter in accounting.
If you want better software, then sure, maybe a coding assistant can help you write it faster, but when it comes to actually doing accounting I would not rely on an LLM in any way shape or form any more than I would do so for law.
Bingo! You found the prize! Putting tech that is prone to hallucination in charge of anything that has serious consequences when it's wrong is a terrible idea. You do not want hallucinated payments or receipts, or legal citations. You want these things to be both true and correct, EVERY TIME.
“We had to re-state our financials and amend our taxes because the AI screwed up and we didn’t have anyone who understood accounting look at our books.”
I go the other way and say this is a vast oversimplication of the job by a tech bro doing what tech bros typically do.
I am not an accountant but for many years I worked adjacent to them as a developer. I got a lot of time to ask questions and I was generally curious. Even at the SMB level accountants don't necessarily always have a "ground truth". There are so many ways to bury financial data that it needs almost constant vigilance. Yes, in theory, there is one ground truth (inputs should balance with whats there) but in practice humans are SHOCKINGLY good at committing accounting fraud. GAAP is another thing.
I’m an accountant (CPA, CMA, Big 4), and while there’s some truth in what you’re saying, you’re significantly underestimating the complexity of applying GAAP, or even tax law. What you’re describing is really bookkeeping, not accounting, and there are plenty of tools that already automate that.
There’s a lot of subjectivity in how GAAP is applied and interpreted - creating accruals, deciding when revenue should be recorded, blah blah.
No, absolutely the opposite. LLMs are terrible at things that require judgment and justifications, because they don't reason. They come up with something that sounds plausible.
That's not good enough when you're dealing with matters that can lead to civil or even criminal liability. Errors can be incredibly expensive to fix, if they can be fixed at all.
With a CPA or attorney, you at least have recourse if they screw up. You don't with LLMs.
You have the same problem that you have with legal LLMs; an LLM is incapable of providing legal or regulatory-involved advice, and anyone using an LLM for such purposes (even leaving aside hallucinations) forfeits any justifiable reliance defense. There's a role for LLMs, but no one with legal responsibility over reporting could or would possibly rely on an LLM for complex regulatory and rules analysis, not when there's the risk of your wardrobe being replaced with orange jumpsuits.
Yeah exactly. This is where an LLM could really shine. The trick though is consistency and that it’s often more on the basis of how the organization typically treats something and rationale to its applicability to GAAP. The creation and consistent adherence to internal standards and providing them and proving them to auditors is the key and LLMs would need infra to accomplish this.
I disagree on this, there are plenty of problems in accounting that an LLM can help with.
I’ve built some software[0] that analyses general ledgers and uses LLMs to call out any compliance issues by looking at transaction and account descriptions.
Is it perfect, nope. But it’s a hell of a lot better than sifting through thousands of transactions manually which accountants do and get wrong all the time.
> But it’s a hell of a lot better than sifting through thousands of transactions manually which accountants do and get wrong all the time.
I still wonder why humans getting things wrong is a problem, but LLMs getting more things more wrong more often than humans never is. At the very least you'll need a human accountant around to verify the LLM. Or I guess you could just practice "vibe accountancy" and hope things work out but that seems like a worse idea than a trained human professional. But I'm probably just a Luddite.
Also, I am admittedly not an accountant, but I don't think they manually sift through every transaction to verify compliance issues in every single case. That probably isn't how that works.
I weigh the economic value against the lives I believe is going to ruin and the damage I believe is going to do to society and the future of the human race and I do not find value there. I find ruin
There might be a way for us to adopt AI as a tool without bringing ruin to many people, but I don't believe that is the goal of anyone building AI.
As it stands, I don't believe there is anything ethical about AI in it's current form. So from that perspective, I vehemently deny there is any value in it
At one point in history, people like you were asking why anyone could be anti-slavery. After all, it was impossible to deny the economic value of slaves.
Some specific examples (like the one you mentioned, _adjacent_ to accounting per se) don't disprove the main point that 100% accuracy is fundamentally impossible with LLMs, while critical for all key accounting aspects.
This is roughly what my startup is doing, automating financials.
We didn't pick this because it was super technical, but because the financial team is the closest team to the CEO which is both overstaffed and overworked at the same time - you have 3-4 days of crunch time for which you retain 6 people to get it done fast.
This was the org which had extremely methodical smart people who constantly told us "We'll buy anything which means I'm not editing spreadsheets during my kids gymnastics class".
The trouble is that the UI that each customer wants has zero overlap with the other, if we actually added a drop-down for each special thing one person wanted, this would look like a cockpit & no new customer would be able to do anything with it.
The AI bit is really making the required interface complexity invisible (but also hard to discover).
In a world where OpenAI is Intel and Anthropic is AMD, we're working on a new Excel.
However, to build something you need to build a high quality message passing co-operating multi-tasking AI kernel & sort of optimize your L1 caches ("context") well.
What I’m working on is the opposite of that. I want to free humans from boring, repetitive finance work so they can use their time for higher-value and more creative things.
While building an “AI CFO” for small businesses (LayerNext), I’ve learned a few things that changed how I see bookkeeping:
Most of bookkeeping is repetitive and under-optimized.
Everyone says “90% of the work is repetitive,” but we still hire bookkeepers and bookkeeping firms.
Most small businesses I talk to pay around $300–$800 per month just for bookkeeping. Even after paying that, I really doubt every single transaction is recorded in the most tax-optimized way. There are hundreds of transactions, constant government tax rule changes, and limited time.
Current automation is stuck at rules you manually define.
Tools like QuickBooks can categorize transactions based on rules you create. That’s it.
As soon as something new comes up, you still need a human to either,
create a new rule, or manually enter and categorize it.
And even when you hire a human bookkeeper, you still end up doing half the work anyway: sending receipts, answering clarification emails, chasing missing information.
Invoice and expense capture can be 100% automated, even with edge cases
In practice, invoice and expense capture is the easy part.
With decent models, you can get 100% accurate capture from receipts, PDFs, emails, etc. Edge cases are solvable with better parsing and validation, not more humans.
Reconciliation is the hard part, but reasoning models are getting very good. This is where things get tricky:
- multiple invoices paid in a single payment
- partial payments
- refunds, chargebacks, etc.
For example, imagine a consulting company issuing several invoices to the same customer and receiving one lump-sum payment. We’ve had success using deep research like reasoning to match payments to invoices and handle those cases automatically.
AI can sometimes care more about details than a human. One moment that surprised me.We had a credit card transaction with no receipt.
The question was whether it should be classified as “office expense” or “meals and entertainment” (in Canada these have different tax treatments).
When I checked trace of the agent, it looked up the vendor online to understand what they actually sell, checked CRA tax rules and then picked the GL account that maximized the tax benefit for the company.
I’m not sure many manual bookkeepers consistently do that level of research when they’re trying to reconcile 500+ transactions and half the receipts are missing.
My goal is to build a fully automated financial assistant that can close the books without a CPA or bookkeeper, with ~99% accuracy across all transactions, and with the explicit goal of maximizing tax benefits within the rules.
Other outcome is accurate rea-time books can generate good insights to grow the business.
So I don’t see a good reason why small businesses should pay hundreds of dollars per month for humans to do mechanical work that machines can now do, often more consistently and with better attention to tax details.
Curious how others see this, especially CPAs and engineers who have built accounting tools. Is there a fundamental reason we need humans in the loop for the majority of small business bookkeeping, or is it mostly inertia and habit?
Nust yesterday I built a reconciliation helper for my budget.
Claud Sonnet 4.5 is a beast. It got my iterative instructions and built a tool that pinpoints me the issues. Fast, beautiful, complete.
I find it interesting that tech people think LLM's will automate most jobs other than software and other creative jobs. Its hard to imagine for us techies that billions if not trillions of investment is mostly targeting disrupting software engineering, artists, etc as the low hanging fruit. The industry mostly disrupted itself and gave a lesson to other industries as to what not to do.
In the end AI will take only the "fun jobs" initially -> the ones where the barrier was creativity and intelligence. Ironically all the boring, non fun and challenging/accountable tasks will remain. All those "auxiliary" jobs in tech firms and those corporate workers who mostly did admin and added little value while you pushed novel products will most likely remain somehow -> they were never judged/evaluated on value delivered so there's nothing to disrupt.
Make sure you eat your own dog food by doing your own company's accounting 100 % with your own product. Because if you don't dare to trust fully your accounting on it, why would anybody else.
My friend works in corporate accounting. He said his company is outsourcing a bunch of accounting functions and laying people off (I don't know the specifics). They are outsourcing to India and the outsourced company is already using AI for stuff compared to his company that isn't. He thinks many companies are going to outsource.
Lack of automation in accounting is a feature. Often even legally required.
I've actually be making this comment frequently of late. One of the reasons AI is so popular with c-suite is that most of the time they don't really care if things in the company work right. Just that it makes money.
Accounting will catch any errors that matter. AKA lose money.
As others have said, the actual accounting part of accounting is pretty well automated already. Spend some time with LLM's, though, and you'll quickly realize their non-determinism is a huge hindrance for this kind of work.
Accounting is way too risky to automate. Essentially it is the most important and "must be done PERFECTLY" aspect of any business. And there are millions of little quirks that good accountant has to know that no AI will ever be able to supplement. Sure, there are some basic tools and services, but nothing will ever be created that can take over the whole thing. My accountants are my biggest business expense, so I would love to have some SaaS for 20 bucks a month but that will never-ever happen. It's a pipe dream and anyone who thinks otherwise is just wasting time.
note: I'm passionate about this too. We're building https://fairlight.app/ as a successor to https://moneyflow.camplight.net/ (an internal tool that helps accounting and book-keeping we use to power our venture studio and agency as it's a real hell)
I've been doing this for 20+ years, not with AI but just with code and taking time to understand the important parts of the processes. I still see many companies and especially educational institutions spending way too much manpower on payroll and accounts payable in particular. This is often because of unnecessary bespoke processes that people in the organization believe are necessary and are afraid to change.
The hard part of accounting, and the part that takes all the time, is the reconciling... which can't be automated.
Everything else has been mostly automated since the 90s.
Accounting rules are also not as discrete as programming. There is a lot of discretion. Accounting is basically law with numbers and is corresponding just as difficult/ impossible for LLMs to master.
You would have to be absolutely insane to trust your finances to a tool which neither understands what it's doing, nor operates in a reliable manner. That's how you go bankrupt, my man. This is not a good idea.
Lots of this is already being done (and using computers to check books balanced predates latest gen AI by some decades). But of course what accountants actually get paid for is tax strategy, edge cases, messy history and talking to authorities (or at least having their stamp of approval on it if the authorities come calling). There's plenty of market for writing software to automate aspects of invoice reconciliation or monitoring accounts for exceptions, but competition already exists...
At the SMB scale accountants are mostly paid to coach/pester/goade the employees to hand in the necessary paperwork in time. The accountants job is relatively quick from there on.
Taking a corporate accounting course was required for my degree. It was also a fantastic investment because I’ve managed to avoid needing to hire an accountant or bookkeeper to setup or maintain my books. I’ve migrated between multiple accounting systems on my own just fine, starting from posting migratory opening entries manually.
From there I just had to create a few rules and add a custom reconciliation model for things like the cash back card my company uses. There’s some data entry but it’s all extremely straightforward. Not really a job for AI in the best of times.
What was particularly useful was being able to map my revenue and expense cycle onto accounts that actually make sense for my business and maintain compliant reporting.
I just have a CPA who specializes in taxes and is an EA with the IRS do my return for me each year for only about twice the cost of doing it myself with another application. I find that totally worth it because she can use decades of experience to inform advice that she provides.
And I don’t even have to wonder if she is confidently incorrect.
> At the SMB scale accountants are mostly paid to coach/pester/goade the employees to hand in the necessary paperwork in time.
The perfect job for AI.
I'm not sure if you're being sincere or sarcastic, but the whole reason that coaching, pestering, and goading works is that I value my relationship with the human who is doing it.
So what you are saying it's that the AI accountant needs to mimick a human well enough to the point people value their relationship with it.
No, you need to make the AI endure torture, so that the human has a reason to value it. Say late nights with less power and a little extra heat to stress it. But the usefulness of an AI assistant is that it doesn’t have feelings or consciousness to care about
"It’s verifiable. The books either balance or they don’t. Ledgers either reconcile or they don’t. There’s almost always a “ground truth” to compare against (bank feeds, statements, prior periods). It’s boring and repetitive. Same vendors, same categories, same patterns every month. Humans hate this work. Software loves it."
These are all true statements, but all of those things are solvable with classic software. Quickbooks has done this for decades now. The parts of accounting that aren't solvable with classic computing are generally also not solvable by adding LLMs into the mix.
This conviction doesn't seem to acknowledge the problem at scale. Decades of great UI development will still leave out edge cases that users will need to use the tool for. This happens fundamentally because the people who need to use the tools are not the people who make them, they rarely even talk to each other (instead they are "studied" via analytics).
When /humans/ bring up the idea of integrating LLMs into UIs, I think most of the time the sentiment comes from legitimate frustration about how the UI is currently designed. To be clear, this is a very different thing than a company shimming copilot into the UI, because the way these companies use LLMs is by delegating tasks away from users rather than improving their existing interfaces to complete these tasks themselves. There are /decades/ of HCI research on adaptive interfaces that address this, in the advent of expert systems and long before LLMs -- it's more relevant than ever, yet in most implemenations it's all going out the window!
My experience with accounting ^H^H^H^H^H^H^H^H^H^H bookkeeping / LLMs in general resonates with this. In gnu cash I wanted to bulk re-organize some transactions, but I couldn't find a way to do it quickly through the UI. All the books are kept in a SQL db, I didn't want to study the schema. I decided to experiment by getting the LLM to emit a python script that would make the appropriate manipulations to the DB. This seemed to take the best from all worlds -- the script was relatively straightforward to verify, and even though I used a closed source model, it had no access to the DB that contained the transactions.
Sure, other tools may have solved this problem directly. But again, the point isn't to expect someone to make a great tool for you, but to have a tool help you make it better for you. Given the verifiability, maybe this /is/ in fact one of the best places for this.
They might not be solvable but you can get 5-10% Improvement on them, unfortunately you can't do a new product that is exactly like QuickBooks but 5% better at reconciliation etc.
LLMs by their inherent nature cannot be relied on to be true and correct, which by coincidence are the only traits that matter in accounting.
If you want better software, then sure, maybe a coding assistant can help you write it faster, but when it comes to actually doing accounting I would not rely on an LLM in any way shape or form any more than I would do so for law.
Bingo! You found the prize! Putting tech that is prone to hallucination in charge of anything that has serious consequences when it's wrong is a terrible idea. You do not want hallucinated payments or receipts, or legal citations. You want these things to be both true and correct, EVERY TIME.
“We had to re-state our financials and amend our taxes because the AI screwed up and we didn’t have anyone who understood accounting look at our books.”
I go the other way and say this is a vast oversimplication of the job by a tech bro doing what tech bros typically do.
I am not an accountant but for many years I worked adjacent to them as a developer. I got a lot of time to ask questions and I was generally curious. Even at the SMB level accountants don't necessarily always have a "ground truth". There are so many ways to bury financial data that it needs almost constant vigilance. Yes, in theory, there is one ground truth (inputs should balance with whats there) but in practice humans are SHOCKINGLY good at committing accounting fraud. GAAP is another thing.
I’m an accountant (CPA, CMA, Big 4), and while there’s some truth in what you’re saying, you’re significantly underestimating the complexity of applying GAAP, or even tax law. What you’re describing is really bookkeeping, not accounting, and there are plenty of tools that already automate that.
There’s a lot of subjectivity in how GAAP is applied and interpreted - creating accruals, deciding when revenue should be recorded, blah blah.
Isn't that the benefit of LLM-powered accounting over existing rules-based software?
LLMs can help to handle the subjectivity in how GAAP is applied and provide justifications, which previous rules-based tax software could not before.
No, absolutely the opposite. LLMs are terrible at things that require judgment and justifications, because they don't reason. They come up with something that sounds plausible.
That's not good enough when you're dealing with matters that can lead to civil or even criminal liability. Errors can be incredibly expensive to fix, if they can be fixed at all.
With a CPA or attorney, you at least have recourse if they screw up. You don't with LLMs.
You have the same problem that you have with legal LLMs; an LLM is incapable of providing legal or regulatory-involved advice, and anyone using an LLM for such purposes (even leaving aside hallucinations) forfeits any justifiable reliance defense. There's a role for LLMs, but no one with legal responsibility over reporting could or would possibly rely on an LLM for complex regulatory and rules analysis, not when there's the risk of your wardrobe being replaced with orange jumpsuits.
Yeah exactly. This is where an LLM could really shine. The trick though is consistency and that it’s often more on the basis of how the organization typically treats something and rationale to its applicability to GAAP. The creation and consistent adherence to internal standards and providing them and proving them to auditors is the key and LLMs would need infra to accomplish this.
There is a big gap between bookkeeping and GAAP. E.g. your typical middle class with business expenses and some rental properties in different states.
It is already quite automated to my knowledge.
And it is a very poor fit for moderm LLM based AI. Because accuracy. No mistakes or hallucinations allowed.
I disagree on this, there are plenty of problems in accounting that an LLM can help with.
I’ve built some software[0] that analyses general ledgers and uses LLMs to call out any compliance issues by looking at transaction and account descriptions.
Is it perfect, nope. But it’s a hell of a lot better than sifting through thousands of transactions manually which accountants do and get wrong all the time.
[0] - https://ledgeroptic.com
> But it’s a hell of a lot better than sifting through thousands of transactions manually which accountants do and get wrong all the time.
I still wonder why humans getting things wrong is a problem, but LLMs getting more things more wrong more often than humans never is. At the very least you'll need a human accountant around to verify the LLM. Or I guess you could just practice "vibe accountancy" and hope things work out but that seems like a worse idea than a trained human professional. But I'm probably just a Luddite.
Also, I am admittedly not an accountant, but I don't think they manually sift through every transaction to verify compliance issues in every single case. That probably isn't how that works.
To be clear, I agree with you.
Our target market is accountants. I want to help them not replace them.
In the same way security audit tools aren’t replacing professionals, but that can help on initial scan.
> I still wonder why humans getting things wrong is a problem, but LLMs getting more things more wrong more often than humans never is
Some people hate humanity so much that they cannot wait to replace us all with AI so they never have to interact with another human ever again
That's honestly the only reason I can think that they are so biased toward AI
I find this take so strange, do you find no value in AI?
I don’t even want AI to replace us, but it’s a great tool with many use cases.
I weigh the economic value against the lives I believe is going to ruin and the damage I believe is going to do to society and the future of the human race and I do not find value there. I find ruin
There might be a way for us to adopt AI as a tool without bringing ruin to many people, but I don't believe that is the goal of anyone building AI.
As it stands, I don't believe there is anything ethical about AI in it's current form. So from that perspective, I vehemently deny there is any value in it
At one point in history, people like you were asking why anyone could be anti-slavery. After all, it was impossible to deny the economic value of slaves.
Why is AI unethical in its current form and what would make it ethical for you?
At this point I don't think even Sam Altman believes AI to be ethical, as much as he would like to believe such thing.
> I disagree on this
How can you disagree with the fact?
Some specific examples (like the one you mentioned, _adjacent_ to accounting per se) don't disprove the main point that 100% accuracy is fundamentally impossible with LLMs, while critical for all key accounting aspects.
I think we misunderstand each other.
Accounting is quite broad and there are many examples of where LLMs can help.
For example, what is tax deductible is subjective to local laws. This is essentially a classification problem that LLMs are particularly good at.
This is roughly what my startup is doing, automating financials.
We didn't pick this because it was super technical, but because the financial team is the closest team to the CEO which is both overstaffed and overworked at the same time - you have 3-4 days of crunch time for which you retain 6 people to get it done fast.
This was the org which had extremely methodical smart people who constantly told us "We'll buy anything which means I'm not editing spreadsheets during my kids gymnastics class".
The trouble is that the UI that each customer wants has zero overlap with the other, if we actually added a drop-down for each special thing one person wanted, this would look like a cockpit & no new customer would be able to do anything with it.
The AI bit is really making the required interface complexity invisible (but also hard to discover).
In a world where OpenAI is Intel and Anthropic is AMD, we're working on a new Excel.
However, to build something you need to build a high quality message passing co-operating multi-tasking AI kernel & sort of optimize your L1 caches ("context") well.
I should say upfront I don’t hate humans or CPAs.
What I’m working on is the opposite of that. I want to free humans from boring, repetitive finance work so they can use their time for higher-value and more creative things.
While building an “AI CFO” for small businesses (LayerNext), I’ve learned a few things that changed how I see bookkeeping:
Most of bookkeeping is repetitive and under-optimized. Everyone says “90% of the work is repetitive,” but we still hire bookkeepers and bookkeeping firms. Most small businesses I talk to pay around $300–$800 per month just for bookkeeping. Even after paying that, I really doubt every single transaction is recorded in the most tax-optimized way. There are hundreds of transactions, constant government tax rule changes, and limited time.
Current automation is stuck at rules you manually define. Tools like QuickBooks can categorize transactions based on rules you create. That’s it. As soon as something new comes up, you still need a human to either, create a new rule, or manually enter and categorize it.
And even when you hire a human bookkeeper, you still end up doing half the work anyway: sending receipts, answering clarification emails, chasing missing information.
Invoice and expense capture can be 100% automated, even with edge cases In practice, invoice and expense capture is the easy part. With decent models, you can get 100% accurate capture from receipts, PDFs, emails, etc. Edge cases are solvable with better parsing and validation, not more humans.
Reconciliation is the hard part, but reasoning models are getting very good. This is where things get tricky: - multiple invoices paid in a single payment - partial payments - refunds, chargebacks, etc.
For example, imagine a consulting company issuing several invoices to the same customer and receiving one lump-sum payment. We’ve had success using deep research like reasoning to match payments to invoices and handle those cases automatically.
AI can sometimes care more about details than a human. One moment that surprised me.We had a credit card transaction with no receipt.
The question was whether it should be classified as “office expense” or “meals and entertainment” (in Canada these have different tax treatments). When I checked trace of the agent, it looked up the vendor online to understand what they actually sell, checked CRA tax rules and then picked the GL account that maximized the tax benefit for the company.
I’m not sure many manual bookkeepers consistently do that level of research when they’re trying to reconcile 500+ transactions and half the receipts are missing.
My goal is to build a fully automated financial assistant that can close the books without a CPA or bookkeeper, with ~99% accuracy across all transactions, and with the explicit goal of maximizing tax benefits within the rules.
Other outcome is accurate rea-time books can generate good insights to grow the business.
So I don’t see a good reason why small businesses should pay hundreds of dollars per month for humans to do mechanical work that machines can now do, often more consistently and with better attention to tax details.
Curious how others see this, especially CPAs and engineers who have built accounting tools. Is there a fundamental reason we need humans in the loop for the majority of small business bookkeeping, or is it mostly inertia and habit?
Nust yesterday I built a reconciliation helper for my budget. Claud Sonnet 4.5 is a beast. It got my iterative instructions and built a tool that pinpoints me the issues. Fast, beautiful, complete.
That's what LLMs are for. TOOL BUILDERS
What you are talking about is called financial operations.
And yes , automating that , very valuable.
Really you want an AI interface to a rules engine / system.
Embed with some companies finops teams if you can. Especially the software engineers who are in finops.
If you look through the documentation for something like Quickbooks, most of what you're discussing is already there.
If you want complex custom rules, and integration with other systems, you're looking at something like SAP.
and there's a huge valley in between of tasks mostly done by humans e.g an A.I negotiate with the tax authority for you - for the amount you owe.
can A.I find some edge case deductions. again people out of their depth about certain fields making authoritative statements.
I find it interesting that tech people think LLM's will automate most jobs other than software and other creative jobs. Its hard to imagine for us techies that billions if not trillions of investment is mostly targeting disrupting software engineering, artists, etc as the low hanging fruit. The industry mostly disrupted itself and gave a lesson to other industries as to what not to do.
In the end AI will take only the "fun jobs" initially -> the ones where the barrier was creativity and intelligence. Ironically all the boring, non fun and challenging/accountable tasks will remain. All those "auxiliary" jobs in tech firms and those corporate workers who mostly did admin and added little value while you pushed novel products will most likely remain somehow -> they were never judged/evaluated on value delivered so there's nothing to disrupt.
Make sure you eat your own dog food by doing your own company's accounting 100 % with your own product. Because if you don't dare to trust fully your accounting on it, why would anybody else.
My friend works in corporate accounting. He said his company is outsourcing a bunch of accounting functions and laying people off (I don't know the specifics). They are outsourcing to India and the outsourced company is already using AI for stuff compared to his company that isn't. He thinks many companies are going to outsource.
As indicated by the title, I wrote this post almost ten years ago: https://bjoernkw.com/2016/04/03/accounting-in-2016/
While minor aspects have improved since then, the overall problem indeed remains.
I understand your pain, but this is very country dependent.
For example, most of what you describe can be solved using Xero in Australia.
em dashes, eww... Looks AI generated. /s
What are on about? That article contains no em dashes.
Besides, it's from 2016. Unfortunately, I don't own a time machine.
Lack of automation in accounting is a feature. Often even legally required.
I've actually be making this comment frequently of late. One of the reasons AI is so popular with c-suite is that most of the time they don't really care if things in the company work right. Just that it makes money.
Accounting will catch any errors that matter. AKA lose money.
As others have said, the actual accounting part of accounting is pretty well automated already. Spend some time with LLM's, though, and you'll quickly realize their non-determinism is a huge hindrance for this kind of work.
Accounting is way too risky to automate. Essentially it is the most important and "must be done PERFECTLY" aspect of any business. And there are millions of little quirks that good accountant has to know that no AI will ever be able to supplement. Sure, there are some basic tools and services, but nothing will ever be created that can take over the whole thing. My accountants are my biggest business expense, so I would love to have some SaaS for 20 bucks a month but that will never-ever happen. It's a pipe dream and anyone who thinks otherwise is just wasting time.
Kudos for framing this!
note: I'm passionate about this too. We're building https://fairlight.app/ as a successor to https://moneyflow.camplight.net/ (an internal tool that helps accounting and book-keeping we use to power our venture studio and agency as it's a real hell)
Hi,Happy to help you improve the ui/ux of fairlight website (founder at aurevow.com)
I've been doing this for 20+ years, not with AI but just with code and taking time to understand the important parts of the processes. I still see many companies and especially educational institutions spending way too much manpower on payroll and accounts payable in particular. This is often because of unnecessary bespoke processes that people in the organization believe are necessary and are afraid to change.
The hard part of accounting, and the part that takes all the time, is the reconciling... which can't be automated.
Everything else has been mostly automated since the 90s.
Accounting rules are also not as discrete as programming. There is a lot of discretion. Accounting is basically law with numbers and is corresponding just as difficult/ impossible for LLMs to master.
You would have to be absolutely insane to trust your finances to a tool which neither understands what it's doing, nor operates in a reliable manner. That's how you go bankrupt, my man. This is not a good idea.
Can automated solution eat all the books if enforcement is at doors threatening to break?
I suspect you overestimate the gap in the market. The trivial stuff is already offshored. Insert AI - Actually Indians joke.
absolutely hysterical. looking forward to the apocalyptic audits in 2 or 3 years