The State of AI in Accounting & Bookkeeping 2026
We surveyed 486 accounting and bookkeeping professionals across North America—and the data shows an undeniable gap between AI implementation and impact.
Introduction
In most small firms, the AI adoption story reads less like a tidy case study and more like a messy diary.
There’s no AI steering committee. No pilot program with clearly defined success criteria. No fancy change-management consultant. Instead, leadership probably said something like, “Hey, I think we should be using this…” And then everybody went off and figured it out on their own, one Claude chat at a time.
In some ways, that approach actually worked. Nearly every firm we surveyed is now somewhere on the AI adoption curve. Only 5% are still sitting it out entirely. And in the grand scheme of things, that is a staggering rate of change for a profession that still raves about how much they love spreadsheets.
The semi-shocking other side of the coin: Only 1 in 5 firms can point to a measurable return on their AI investment. (Though about half say the value of AI appears promising; they just can’t quantify it yet.)
And the single biggest thing standing between accounting firms and true ROI isn’t cost, or fear, or access to the technology itself. It’s that nobody has an hour to spare. 😅
If that sounds familiar, then you’re in good company. And we’re here to help.
We believe the path to improvement starts with education. So, to get a better understanding of where the industry actually stands with AI in accounting, we surveyed 486 bookkeeping and accounting professionals about:
- How far along they really are,
- What they’re using AI for (and what they aren’t),
- Whether any of it is paying off, and
- What they think it all means for the future of the profession.
What follows is an unfiltered look at the state of AI in small and growing accounting firms: the adoption landscape, the ambiguity around impact, and the clear need for better infrastructure.
Survey Methodology
Data collection:
Online survey conducted by Financial Cents from July 15–August 7, 2026
Median completion time = 10 minutes
- Data on the profession as a whole (i.e., adoption stage, barriers, hopes, worries, predictions, etc.) based on answers from all 486 respondents. (This is the Full Base.)
- Data about hands-on use (i.e., tools, tasks, time saved, ROI, policy, governance, etc.) based on answers from the ~380 respondents at firms already using AI. (This is the AI-Active Base.)
Who Responded: A Quick Look at the Firms Behind the Data
Before we get into the good stuff, let’s meet the people behind the numbers. (Spoiler: They probably look a lot like you. 😉)
What’s your role at the firm?
Full Base
How many people work at your firm?
Full Base
How many years have you worked in the profession?
Full Base
- < 2 years — 11%
- 2–5 years — 12%
- 6–10 years — 16%
- 11–20 years — 24%
- 20+ years — 37%
Which services does your firm offer? (Select all that apply)
Full Base
Survey Base Snapshot
are owners or partners
work in firms of 2–30 people
have 11+ years’ experience
Two-thirds of respondents are firm owners, and 3 in 5 have worked in accounting for more than a decade.
Where Accounting Firms Fall on the AI Adoption Curve
How far along is the accounting profession, really? Further than you might guess, but not quite as far as the headlines suggest. Here’s a detailed picture of industry-wide AI adoption.
Where would you put your firm on the AI adoption curve?
95% of firms have adopted AI in some capacity, but only 11% are “running” with it.
What’s the biggest barrier to getting more value from AI at your firm?
Base: Non-Active AI Users
trust in outputs
& client confidentiality
where to start
learn or implement
isn’t convinced
uncertainty
Cost is the least common reason firms haven’t gotten started with AI, coming in dead last at 3%. Instead, the two biggest blockers for non-users are trust and security. Overwhelm plays a role as well, with a combined 28% saying they either don’t know where to start or don’t have time to learn/implement the technology.
Straight from the Source: What’s Holding Us Back
Surveytakers explain, in their own words, what’s keeping them from getting more out of AI.
Things are evolving SO FAST that it is almost impossible to keep up. It would be a full-time job.
Usually, if I have a task that I think can be improved by AI, I will try and do it with AI first. So usually 1–2 times a day I will do a task with AI, and some tasks are still not possible.
I wish I had more time for evaluation and that there was a resource on how to implement AI in specific scenarios.
Need to learn more and implement more soon in order to keep up.
In Canada, the CPA regulatory bodies are resistant to AI adoption, so I’m waiting for clear guidelines.
How Firms Are Actually Using AI
This section surprised us the most, and we think it’ll surprise you too. The way accounting professionals actually use AI doesn’t line up with the hype (or with where the biggest efficiency gains are supposed to be).
Which AI tools does your firm use? (Select all that apply)
AI-Active Base
Nearly every firm using AI has adopted a general-purpose chatbot.
Only 1 in 5 are using a tool built for bookkeeping.
General-purpose assistants dominate current AI use, but firms are also beginning to adopt AI inside the systems where work already happens. Financial Cents’ AI features for accounting firms support workflow creation, client communication, file validation, and file naming inside the practice management platform.
Which tasks are you personally using AI for?
AI-Active Base
The top 3 uses are all related to communication and research. Bank reconciliation and tax return prep sit at the bottom of the list.
Furthermore, 24% of AI users aren’t using AI for any core accounting tasks. Instead, they are leveraging AI exclusively for comms drafting, data summarization, research, documentation, and marketing.
Straight from the Source: The Most Valuable Thing AI Does for Us
Surveytakers explain how AI benefits them the most.
Turn PDF bank statements into usable CSV files.
Automate manual processes that don’t directly add to client value.
In the same amount of time, provide better service.
Work more projects in less time. However, the key is you still have to understand the accounting concepts and double-check the work.
Catch errors before they reach a client deliverable—through mandatory verification of every AI-produced figure against source, backed by a human review gate.
It helps us check our work for accuracy, which is very important in the tax/accounting world.
The Payoff Problem
Okay, so (almost) everyone in accounting is using AI. But is it actually working?
Sort of? Maybe? Let’s just say most folks aren’t seeing any earth-shattering results quite yet.
Roughly how much time does AI save you personally each week?
AI-Active Base
More than half of AI users save 3 hours a week or less. Only 15% save 7 hours or more.
Has your firm seen a clear return on its AI investment?
AI-Active Base
Only 4% of adopters say AI hasn’t paid off. But only 20% can prove it has.
The dominant ROI experience is one of ambiguity, not disappointment. (Which is arguably worse, because at least you can act on disappointment.)
The Key to Unlocking ROI = Deepening Adoption
When we looked at differences in measurable ROI and time savings by adoption stage, a clear pattern emerged.
Firms that are “running” with AI across the whole organization are more than 15 times as likely to report measurable ROI as firms where a few people dabble. That means adoption alone isn’t enough to reap the full rewards of AI. You have to commit to leveraging it to the fullest first.
That doesn’t happen on its own. To move beyond the “crawling” stage, you must be intentional about weaving AI into your operations. Informal, optional, individual AI use isn’t enough to produce a measurable return.
2 Keys to Achieving Measurable ROI
1. Structured skill-building: 27% of firms that are doing formal training, designating an internal champion, or using vendor training report clear ROI, compared to 12% of firms that aren’t.
2. A written AI policy: 35% of firms with a formal documented AI policy report clear ROI, compared to 14% of firms with no policy and no plans to create one.
The Trust Ceiling
Here’s where things get interesting. Underneath all the noise, accountants and bookkeepers have arrived at a fairly consistent position on AI, and it’s probably not as negative as you think.
Overall, how do you feel about AI’s role in the profession?
AI-Active Base
- Very optimistic — 33%
- Cautiously optimistic — 39%
- Mixed / neutral — 18%
- Somewhat concerned — 6%
- Very concerned — 3%
Over 7 in 10 accounting professionals are optimistic about AI, while less than 1 in 10 are concerned. Whatever’s slowing down adoption, it isn’t dread.
How much do you agree with each statement?
AI-Active Base
Captured using a 5-point scale (here, “agree” combines “agree” and “strongly agree,” and “disagree” combines “disagree” and “strongly disagree”)
3 Key Takeaways on AI Sentiment
1. 90% agree human judgment matters more, not less, with the advent of AI. This is the most agreed-upon statement in the entire survey, with a near-unanimous vote across every role, firm size, and adoption stage. If the profession has a shared consensus about AI, this is it.
2. Only 19% trust AI enough to use it with limited review. This is even true for firms describing their adoption stage as “running” across the whole organization. Human review is not seen as something firms will eventually grow out of. Rather, it is part of the design itself.
3. Only 18% say clients expect them to use AI. Firms are not being pushed into AI by their clients. They’re choosing it for themselves. This lack of external pressure is a good thing, because it means firms still get to decide what “good” looks like before the market decides for them.
Straight from the Source: What Worries Us Most
We asked surveytakers what worries them most about AI in the accounting profession, and we received 463 unique freeform answers. Coded across all of them, the dominant themes were:
Here’s a sampling of what respondents said about why they’re staying AI-cautious.
AI can produce answers that sound extremely confident, but confidence doesn’t always equal accuracy.
Just because it confidently did the task correctly one time doesn’t mean it won’t do it confidently wrong the next.
The key thing to never forget: the liability always stays with the human.
Lack of critical thinking. Equivalent of, ‘If your GPS tells you to drive into a lake, do you drive into the lake?’ Just because your GPS is on does not mean that you shut your brain off.
It’s the difference between just using a calculator to get an answer and understanding how the calculator came up with the answer in the first place. How can you verify if something is right (or wrong) if you don’t understand the basics of the concepts?
I’m worried that if AI is wrong, there is nobody to hold accountable for the mistake that it made.
People tend to blindly trust it without verifying the output is correct.
I use it to make me more efficient, not replace me.
Who will train the next generation?
Roughly 5% of the responses about AI worries spoke to a common concern: AI is absorbing the work that junior staff used to learn on. (Things like transaction coding, data entry, and tedious first passes.)
I worry that giving the day-to-day coding of transactions or more tedious work to AI means there are less entry-level positions and no one will be able to learn the basics. This would create a large gap between legacy professionals and new professionals who don’t have the same solid foundation.
It feels like AI will reduce/eliminate staff level work. My concern is that a how I learned my job was a staff level in the trenches. I worry those following me won’t be ready. It’s a BIG step up.
There needs to be better accounting and tax knowledge transfer in order for lower level staff to enter and be proper experience to oversee AI.
One owner with more than 20 years in the profession drew the line straight back to their own training:
Making sure we do not lose the entry-level pipeline. Working with all my AI vendors to make tools available to students so they will learn how to supervise while they are getting training/experience in the doing—just like I learned on green ledger paper during the ’80s when computerization was making transaction entry “automated.” I still use T-accounts to map my journal entries and adjustments to make sure my computer is correct!
This concern grows even more troublesome considering that 90% of the profession says human judgment matters more as AI spreads, and 19% trust AI output without close review. Both of those positions depend on having people who know enough to catch the mistakes, and the traditional way people developed that ability is through the exact work many firms are offloading first.
The Governance Gap
Time for the uncomfortable section. Deep breath. 😬
Does your firm have a written AI policy or other internal usage guidelines?
AI-Active Base
- Yes, we have a formal, documented policy — 13%
- We have informal guidelines / verbal norms — 12%
- We’re in the process of developing this — 18%
- No, but we plan to — 29%
- No, and we don’t have any current plans to — 23%
- Not sure — 5%
Of the 350+ firms actively using AI, 87% have no formal written AI policy, and 23% have no plans to create one. That means 9 in 10 firms are putting client work through AI tools with no documented rules about what can and can’t be entered.
This is a particularly concerning problem for the industry’s smallest firms: 37% of solo practitioners have no policy and no plans to create one. Solo owners are exposed twice: no documented rules, and nobody else to catch any issues.
So how are firms handling client data?
We asked this as an open question and got 359 answers. Coded across all of them, six patterns emerged.
Straight from the Source: How We Handle Client Data
If you’re building an AI policy from scratch, this is a great source of inspo.
I treat AI as an additional service provider that must meet the same confidentiality standards as any bookkeeping platform. I use only approved business-grade tools with appropriate security, privacy, retention, and ‘no training on client data’ protections.
Treat AI as a brand-new employee that knows nothing about your business. It needs all the safeguards and limits that you give a new employee.
We use business plans that have agreements not to save our prompts and data. When not possible we redact information before feeding into AI. We also have a term in all of our agreements stating the client is aware we use AI and there are inherent risks that they are okay with.
We know which clients do not wish us to use it at all. We have engagement authorities permitting use with clients. Client data is not put into open channels. Where an open channel is being used, all data is anonymized prior to use.
Nothing with client names or identifying details is ever uploaded to an external AI source.
Be careful with AI recorders. [Our state] is a two-party state, meaning you need consent to record.
The governance gap is part of a broader technology-management problem. The 2026 Bookkeeping Firm Tech Stack Report found that many bookkeeping firms still rely on disconnected tools and manual data entry, making it harder to maintain visibility and consistent controls.
How Firms Are Learning AI
We’ve established that depth pays and structure helps. So why doesn’t every firm just…do that?
What’s the biggest barrier to getting more value from AI at your firm?
Full Base
Across our full survey base, time is the biggest barrier to getting more value out of AI, by double. In fact, it’s more than twice as big as cost, tool overload, and team resistance combined.
Even more telling: “unclear ROI” ranks dead last at 2%. Firms aren’t holding back because they doubt AI will pay off. They believe it will. They just haven’t had enough time to prove it out.
How is your firm building AI skills? (Select all that apply.)
AI-Active Base
informally
structured yet
/ courses / CPE
software-provided training
a designated lead
AI skills
How is AI usage spreading through your firm?
AI-Active Base
- Top-down: leadership is driving it — 52%
- A mix of both — 31%
- It isn’t really spreading yet — 13%
- Bottom-up: individuals adopting on their own — 4%
This runs counter to the standard story about AI in the workplace. In small accounting firms, AI adoption isn’t a grassroots staff movement. It’s leadership-led in more than half of firms.
Straight from the Source: Advice to Firms Just Getting Started with AI
When asked what advice they would give another firm just starting out with AI, 353 survey-takers offered up their words of wisdom. The biggest themes we found across their responses:
Here’s a sampling of direct quotes about implementation best practices.
Start small and focus on solving one real problem at a time rather than trying to implement AI everywhere.
Create a basic AI policy before expanding: use approved business-grade tools, minimize confidential data, verify every output, and keep a human responsible for the final work.
Do not use AI for things you don’t know how to solve. AI is only as good as the data it is given and the question being asked.
Don’t be afraid to abandon it for a specific task and go back to the old way. Clients are paying you to do work they trust, not to be the most technologically advanced.
Do a personal audit of your skills and weaknesses, especially the areas you procrastinate or really don’t feel effective, and then determine if AI can effectively support those areas.
Use the time AI saves to become more responsive and advisory. The goal isn’t simply to complete the same bookkeeping faster; it’s to give clients more clarity, attention, and confidence.
Role Spotlights: Same Technology, Different Jobs
A lot of existing AI research treats each accounting firm like one unit that either adopts or doesn’t. But that’s not really how things work, especially as teams grow.
Different roles have different AI use cases, available tools, and risks. So, we created a unique set of questions for each role type—and uncovered a wide range of priorities, use cases, and anxieties.
A note on the numbers below: Some role-based groups had greater representation than others. For groups with smaller sample sizes, we have presented numbers as exact response counts rather than percentages.
8A · Firm Owners & Partners
330 respondents
As an owner, what’s your top priority for AI right now?
Owners want greater capacity, not bigger margins. Profitability ranks as their fifth priority, just behind “staying competitive.” The dominant goal for owners is scaling up work without hiring more people, which lines up neatly with the 52% who agree AI lets their firm do exactly that.
The pricing conversation owners are having with themselves
An interesting clash that came through in the data: Only 4% of all respondents selected “pricing will move away from hourly billing” as the biggest change coming by 2030. But when we gave owners an open text box to talk about growth and pricing, billing methodology was one of the most common themes they referenced.
As routine work becomes more efficient, hourly pricing makes less sense because it can penalize the firm for improving its systems. I am moving toward value-based, fixed-fee packages that reflect the outcome, level of support, complexity, and responsibility involved.
AI isn’t changing what I sell. It is changing how I deliver it. I don’t think AI creates a sustainable competitive advantage anymore because everyone has access to it. The advantage comes from working with a business, judgment, and execution.
We want to remain premium priced, which means we have to deliver higher-level value. People don’t want to pay for bookkeeping when AI will increasingly do that for them.
How AI is changing hiring and resourcing
While many owners are thinking about hiring less, others are thinking about hiring differently (with a particular emphasis on being simultaneously AI-forward and AI-skeptical).
My first thought is how I can automate before I even think about training someone. I now hire my staff with a different skill set. Yes they need to know the bookkeeping and accounting rules, but they have to embrace and be curious around technology. I need them to have a good level of doubt the computer is giving them the right answer.
All of that said, not every owner is on board with AI yet. And in some cases, they are actually rolling back their AI investments in order to simplify operations.
I’m not a fan of AI at all. I know as there are a lot of mistakes and prompting needed. It causes more issues than not. We’ve been actively reducing it from our firm.
8B · Service Line Leaders
92 respondents (senior accountants, senior bookkeepers, tax and advisory leads, and staff accountants)
In your service line, how do you use AI relative to your own review?
- Mainly to research / sanity-check my own work — 41%
- AI generates drafts, and I review everything closely — 23%
- I rely on it for ideas, but I do the work myself — 17%
- I don’t use it in my service line yet — 14%
- AI handles routine work, and I spot-check — 5%
Just 5% of service line leaders let AI handle routine work
with nothing more than a spot-check.
For service line leaders, the dominant approach to AI is using it as a second set of eyes rather than relying on it for the first pass. Combine that with the 22% who are actively concerned about AI in the profession and the 32% who cited trust and accuracy as their biggest adoption barrier, and a coherent picture comes into focus: the people doing the technical work of accounting think of AI mainly as an assistant requiring close oversight.
How service line leaders are leaning on AI as a support
I use it as a support tool for work more efficiently, while I still review and verify the information before applying it in client work.
I use it to help with transaction categorization in QuickBooks Online. If I meet with a challenge or something new to me, I will often use AI to help become more knowledgeable.
I usually use ChatGPT to help me think something through when I need to bounce ideas off someone else.
8C · Operations & Practice Leaders
23 respondents
Of the 23 operations leaders we surveyed, 17 named standardizing processes and SOPs as the place where AI is having the biggest operational impact, and 16 named workflow and job management. Far lower on the list: 12 named tool and tech-stack decisions.
How AI is fundamentally changing ops work in accounting firms
Honestly, it’s flipped my whole job. I used to be heads-down doing the actual review work—categorizing, reconciling, digging through documents. Now AI does the first pass, and I’m the person who built the system that does that pass, and my job is making sure it’s not lying to me. It’s not ‘set it and forget it.’ It’s more ‘trust but verify,’ always. Less grunt work, more judgment calls.
I have a billing rundown every day that tells me what needs to be billed or followed up on, and a follow-up digest at the end of the day that loops back through all of the outstanding requests. I have a notebook, but my notebook doesn’t remind me in an hour. I use it to make me more efficient, not replace me.
8D · Admins & Firm Administrators
29 respondents
Of the 29 admin folks we surveyed, 21 said email and communication drafting is what AI changed most for them. Beyond that, 14 named meeting notes and follow-ups, and 12 each named document collection and reporting or data pulls.
How admins are streamlining client communication and data collection with AI
Fine-tuning engagement letters, welcome letters/emails, templates that involve communicating with clients.
Text and email responses. Now they are generated for me.
AI has helped with data pulls for credit cards and bank accounts that do not have a feed in Xero.
On the other end of the spectrum, admins were also the most likely role to push back on AI use, largely due to concerns over client data security.
Nothing. AI is largely inefficient, and client data is too sensitive to be entering into an anonymized AI platform that has been known to spit out people’s sensitive data to other unknown users.
There is nothing that used to eat up my day that AI now handles.
What Comes Next
We ended the survey by asking people to look up from their desks and look forward to the AI-supported future. Two things stood out: the timeline is shorter than most firms are planning for, and the skills people expect to matter might not be the ones you’d guess.
When will agentic AI be doing meaningful firm work?
Full Base
(“Agentic” = tools that take actions on their own, not just generate output.)
our firm
this means
Two-thirds of accounting professionals expect agentic AI to be doing meaningful work in their firms within 3 years.
1 in 5 say it already is.
Now, going back to our “Crawling,” “Walking,” and “Running” groups, half of those who are “Running” say agentic AI is already doing meaningful work at their firm (compared to only 9% of those who are “Crawling”).
Additionally, 8% of respondents told us they don’t know what “agentic” means, suggesting that the vocabulary is a bit ahead of the profession.
What’s the biggest way AI changes accounting firms by 2030?
Full Base
- Most compliance / bookkeeping work will be fully automated — 34%
- Smaller teams will be serving far more clients — 20%
- New roles will emerge (AI oversight, data management, prompt design) — 18%
- Firms will shift primarily to advisory & strategy work — 17%
- Pricing will move away from hourly billing — 4%
- Not much will change fundamentally — 4%
A third of respondents expect compliance work to be fully automated by 2030.
Only 4% think not much will change.
What skills will matter most as AI advances? (Select up to 3.)
Full Base
With respect to the skills that will matter most in the AI era, critical thinking and client relationships beat AI fluency roughly 2:1.
The consensus: Judgment, not technical skill, will be the biggest professional differentiator as we move deeper into the AI future. (And remember, 90% of respondents also told us human judgment matters more, not less, in the age of AI.)
It all comes down to automating the pure production work, but keeping a strong human grip on all the things a machine can’t be held accountable for.
The TL;DR
Out of 486 bookkeeping and accounting professionals who completed the survey:
- 95% are on the AI adoption curve, but only 11% are “running” with it across the entire firm.
- Only 1 in 5 can point to measurable ROI. Half say it’s promising but hard to quantify.
- 87% of firms using AI have no written AI policy. Nearly a quarter have no plans to create one.
- The top 3 most common uses are email drafting, document summarization, and research. Bank reconciliation and tax prep are the least-reported uses.
- 90% say human judgment matters more, not less in the AI era—the most agreed-upon statement in the entire survey.
- Only 19% trust AI enough to use it with limited review.
- Only 18% say clients expect them to use AI, but more than half say firm usage is being driven by top-down pressure from leadership.
- Time is the biggest AI adoption barrier, with 41% of the vote (more than double the next answer). Unclear ROI ranks last, at 2%.
- Firms running AI firm-wide are more than 15x more likely to report clear ROI than firms where only a few people dabble.
Contrary to assumption, small firms aren’t avoiding AI out of reluctance or resistance. In fact, they’ve adopted AI faster than almost anyone predicted, mostly by their own initiative—without a budget, a mandate, or any other type of external influence.
But, while most firms aren’t AI-resistant, they also aren’t AI-proactive. Few have built the foundational structures, systems, and policies that lead to optimal AI use. In the years to come, that’s what will separate the “running” firms from those that are still “crawling.”
And wherever you are in your AI journey, Financial Cents will be here to support you every step of the way.
About Financial Cents
Financial Cents is an accounting practice management software built for small and growing firms. It brings workflow management, client communication, billing, document management, and AI-powered tools into one platform, so firm owners can stop stitching together disconnected systems and focus on serving their clients. Over X,000 firms use Financial Cents to manage their practice.
On average, firms save 56 hours per month & $19,200 every year with Financial Cents
Book your free demo today. Book a demo