Willingness to adopt new technology is not what holds firms back. In the AICPA's 2025 National MAP Survey, 88% of responding firms were confident or not concerned about their ability to adapt to AI and automation over the next three years, and most had not yet engaged in formal programs or begun to take full advantage of the technology available to them. Accounting technology adoption lives in the space between those two facts. Three questions close it: who owns the rollout, how you prove it happened, and how you tell a stalled tool apart from a purchase in the wrong layer or a capacity gap wearing a software costume.
What Accounting Technology Adoption Means Inside a Firm
Adoption is not the license and it is not the login. A tool is adopted when one named job type runs on it every time, for every client who gets that job, with the old way of doing it closed.
That definition matters because the survey shows how narrow real use can be while still counting as use. Almost half of firms, 47%, were using new technologies to help with client communications or recommendations, while 13% used them to support real-time dashboards or business insights and 12% to automate cash flow forecasting, tax planning, or scenario modeling (2025 National MAP Survey Executive Summary). One use case is a real start. It is not a job type that has moved.
How ready firms are depends on which question you read. In that survey, 88% of responding firms were confident or not concerned about their ability to adapt to AI and automation over the next three years, while most had not yet engaged in formal programs or begun to take full advantage of new technologies. Asked the narrower question, whether they were confident about adopting AI and automation over those three years, 56% of firms overall said somewhat or very confident (2025 National MAP Survey Executive Summary). Not concerned is not the same as ready, and neither number is a measure of use.
Which product belongs in which part of the stack is a different decision, and the layer question is worked through in choosing accounting software by layer. Everything below assumes the purchase is already made.
Someone Has to Own the Rollout, With Hours Attached
Adoption stalls for want of an owner far more often than for want of enthusiasm.
When the 2025 National MAP Survey asked what was blocking implementation of emerging technologies, lack of time to explore or implement was the top answer at 41%. The distant second choices were lack of internal expertise or leadership, uncertainty about where to start, and adoption already being under way, while staff resistance or fear of change came in at 6%, which the survey noted hardly registered as a deterrent (2025 National MAP Survey Executive Summary). Attitude is not the barrier either: 65% of firms called themselves proactive or open-minded about AI, 24% remained cautious, and 10% were resistant or not yet having the conversation (2025 National MAP Survey Executive Summary).
So an owner is a person, a date, and hours that nobody is billing. CPA Australia commissioned research into the same problem, interviewing 20 managers, senior managers or proprietors in accounting roles at Australian organisations between July and September 2021 about the strategies they used and what got in their way. On time, the report is blunt: staff responsible for technology adoption were time-poor, or the timeline for adoption itself created significant challenges (Strategies for Managing Barriers to Technology Adoption). Those interviews are now five years old, so read the methods rather than the market.
The same interviews put a shape on ownership. One manager described building a champions group of people drawn from different work areas who stayed involved through the stages of the rollout, becoming "the flagbearer of the product in the early stages". Another said champions have to be trained intensively first, so they understand how the tool integrates, can troubleshoot what comes up, and can feed problems back (Strategies for Managing Barriers to Technology Adoption).
For the project team itself, one interviewee named the manager of the area where the first tests run, the CFO, the IT manager and an accountant as the right mix, because between them the team knows the systems and the money (Strategies for Managing Barriers to Technology Adoption). A small firm will not have four such people. It still needs the four viewpoints, even if two of them are the same partner.
Rollout Work Is Training, a Sequence, and a Closed Old Path
Seats become use through three specific pieces of work, and none of them are the vendor's job to finish for you.
Training that continues past go-live. The CPA Australia research found staff need ongoing and continued support when adopting new technology, not just support before implementation. Bite-size and practical training, such as an instructional video followed by time to practice on the tool, was described as the optimal approach and preferred over multi-day workshops without enough time for authentic practice. External experts, often the software vendor, are useful for initial training, and once internal team leads and champions are trained they can take it over, which the report calls cheaper and more effective because they can tailor it to what staff actually need (Strategies for Managing Barriers to Technology Adoption).
A sequence that fits the firm's size. In the same study, smaller organizations suggested a fast and full rollout, while medium and larger ones emphasized progressive and staged deployment, or a pilot site followed by full deployment (Strategies for Managing Barriers to Technology Adoption). A ten-person firm that stages a rollout across four quarters is mostly buying itself a year of running two processes at once.
A date the old path closes. Running the new system in parallel with the existing one until the rollout is ready was one of the internal supports the study named. Parallel running is sensible and it has a cost nobody schedules, because while both paths are open, whoever is under the most pressure keeps using the one they already know. Decide the closing date when you open the new path, and record it where the job record lives (what accounting workflow software has to represent).
Say a firm licenses a client document-request tool before extension season. The rollout scope is one job type, individual returns for business owners, and the whole exercise is judged on how many of those engagements collected their documents through the tool rather than by email.
How to Tell Whether Adoption Actually Happened
Adoption is a fraction, and the denominator has to be chosen before the tool goes live.
One survey finding gets quoted as though it were a diagnosis. In the 2025 National MAP Survey, 40% of firms said they had not yet figured out how to track efficiencies due to technological advancements (2025 National MAP Survey Executive Summary). That sits alongside the spending and pricing picture, and what the freed hour is for, in the challenges CPA firms face this year. What is missing is rarely a dashboard. It is a unit.
The Adoption Rate Is a Share of Jobs, Not a Count of Logins
Name the job type. Count how many instances of it the firm ran in a period, then count how many of those ran end to end on the new tool. That share is the adoption rate. It comes out of the job record rather than out of the software's own usage screen, because a vendor's counter measures activity inside its product, not whether the job finished there.
Take the Baseline While the Old Process Is Still Running
Cycle time, meaning the elapsed days from a job starting to it being delivered, is countable today. So is the number of times a file changes hands, and the number of points where work goes back for rework. A baseline reconstructed after the fact is a memory, and memories get argued with in exactly the meeting where the number matters.
Measure the Effect Separately From the Use
A job type can move completely and change nothing. The CPA Australia interviewees evaluated their rollouts on improvements in efficiency, accuracy or productivity in the processes the technology touched, including tasks completed with fewer errors, in less time, or with increased outputs, plus compliance with the new process and statistics on client and staff adoption. They also gathered staff feedback on ease of use and perceived usefulness. Only one of the 20 managers interviewed said their organization used no formal evaluation process at all (Strategies for Managing Barriers to Technology Adoption).
Do Not Use Utilization as Your Adoption Metric
Utilization is chargeable hours divided by hours worked, and it is one of the KPIs worth tracking for other reasons. It is the wrong instrument here, because a tool that takes hours out of a job lowers the numerator. A rollout that works can read as a decline.
The survey's own takeaways make the point from the other direction. Utilization and chargeable hours have declined for entry-level staff with one to three years of experience, which the survey says suggests several issues, including a lack of clearly set goals and expectations as well as a failure to delegate appropriately or to upskill (2025 National MAP Survey Executive Summary). None of those readings are about a tool.
Tracking is itself a habit of the more profitable firms. The survey defines top performers as the top 25% of firms by net remaining per partner, which is net client fees minus expenses before partner compensation is taken out, and one of the differentiators it lists for them is that they are better at tracking efficiencies related to technology advances (2025 National MAP Survey Executive Summary). The survey notes many of these differentiators may be attributable to firm size, so read the tracking habit as a pattern among more profitable firms rather than as what made them profitable.
Three Reasons a Tool Stalls, and the Order to Test Them
A tool that is not paying for itself has three common explanations, and they need testing in order, because each one produces different evidence and only the first is fixed by trying harder.
First, Did the Job Type Ever Fully Move?
Look at the adoption rate. If it never reached the whole denominator, this is an adoption failure, and the tell is a split rather than a flat line: some preparers on the new path and some on the old, or one office on it and one not. Nothing about the tool is disproven yet, because it has not been tested on the work. The fix is the ownership, training and closing date above.
Second, Did the Tool Remove Any Touches?
If the job type did move and the baseline did not improve, stop training people. The tool is not doing work the job actually needs, which is usually a question of what it was built to hold rather than a question of skill. The systems a firm buys each own a different record, and a product bought at the wrong layer gets used correctly and still changes nothing (who owns the job record). More training on a wrong-layer purchase buys a firm nothing except compliance with its own mistake.
Third, Is the Constraint Somewhere the Tool Never Reached?
If real hours came out of the job and the same weeks still break, the constraint sits outside what you moved. Do the subtraction explicitly: compare the hours the rollout removed against the gap the plan says the firm has, using the arithmetic in building a capacity plan that holds. If removing the same hours again would not close the gap, more software will not either, and the shortfall is people rather than features.
Questions Firms Ask About Accounting Technology Adoption
What Does Technology Adoption Mean?
In a firm, it means a job type has moved onto a tool completely, for every client who gets that job, with the previous method retired. Buying, installing and assigning seats are earlier steps that can all finish while adoption sits at zero.
What Are Some Examples of Accounting Technologies?
The systems split by the record each one owns. A general ledger holds one set of books. Tax compliance software is where a return is prepared and transmitted. Practice management tracks the job itself, meaning whose desk it is on and what is due next. Document management holds the workpapers and the retention rule. A client portal is the exchange, the way paperwork goes out to a client and comes back. Naming the layer before naming the product is the part firms skip, and choosing accounting software by layer works through what each one is responsible for.
How Long Should Adoption Take?
It is better read as a share than a duration, because a rollout is finished when the denominator is covered, not when the calendar says so. Set the closing date for the old path at the start, keep the checkpoints short, and treat a rollout still running two seasons later as evidence that the scope was too wide rather than that the firm is slow.
Is Software or Extra Capacity the Better Answer?
That depends on which of the three failures your evidence points to, and the comparison of what automation and added people each solve is worked through in AI versus offshoring in accounting.
Start With One Job Type and a Number You Wrote Down First
Getting value out of technology has less to do with picking the right product than with what happens in the months after the invoice. A named person has hours to run the rollout. One job type moves completely. Somebody writes the before number down while the old process is still running. Anything short of those three is a license with a login page.
Pick one job type this month. Write down its cycle time and its rework points before anything changes, name the owner, and set the date the old path closes.
If the answer turns out to be capacity rather than software, the question becomes who does the work. Accountably places trained offshore accountants and tax preparers inside US CPA and EA firms in about 3 to 4 weeks, ramped on your software and your own written procedures, and a firm can grade the output on a Free 40-Hour Proof Pilot of its own representative work before a client file is at stake. Don't trust us. Test us. Start here.