Executive summary
A practical look at AI change management for professional firms. Computers took years to settle into Australian businesses, and they paid off best alongside skilled people and reorganised work. We set out what that history suggests for AI, what success can look like for a team, and five small changes to start with.
Computers took a long time to reach most Australian businesses. In 1993-94, only 49% of Australian employing businesses used a computer, according to the Australian Bureau of Statistics. Productivity Commission research found that ICT delivered stronger productivity gains when it was paired with skilled people, advanced business practices and organisational restructuring.
In many firms, people have taught themselves AI tools on their own initiative, and fitted the learning in around client work.
You don't have to change everything right now. What we tend to see is that the changes that stick are small ones, made by a team together.
1Where AI sits today
People usually start with tasks they already do. They ask AI to tidy an email, summarise a long document or check a column of figures.
The Reserve Bank asked 105 firms in its liaison program about AI in mid-2025. Its November 2025 Bulletin reported that "adoption has been shallow to date, with nearly 40 per cent indicating minimal use so far". For those firms, use was typically limited to digital assistants such as Microsoft Copilot or ChatGPT. Around 30% had made "more substantive progress".
In the 2025 Victorian Lawyer Census, run by the Victorian Legal Services Board and Commissioner, 37% of 1,739 respondents used AI tools in their practice, rising to 46.0% of those working in law firms. The report found AI's benefits "both narrowly conceived and unevenly distributed", with uses that could improve service quality and access to justice still "largely unexplored".
In Microsoft's 2026 Work Trend Index, a vendor survey of 2,000 full-time Australian workers who use AI, 51% said "it feels safer to focus on current goals than rethink work with AI."
Over time, the firm can start working out new workflows: which steps a piece of work needs, who does each one, and where AI fits.
2Lessons from the computer years
In the same Productivity Commission study, by Paul Gretton, Jyothi Gali and Dean Parham, the first impact of taking up computers tended to be small, and the benefits grew over about five years of use.
In the United States, Erik Brynjolfsson and Lorin Hitt concluded in 2000 that "organizational 'investments' have a large influence on the value of IT investments." Two decades later, Brynjolfsson and his colleagues argued that general purpose technologies, AI among them, both enable and require significant complementary investments. We tell the longer story of how earlier technologies changed working life in Value, time and productivity.
In a study led by Phillippa Lally at University College London, the average time to form a new daily habit was 66 days, and the range was 18 to 254 days. Missing one opportunity "did not significantly impact the habit formation process", UCL reported.
In the Victorian census, 72.3% of lawyers using AI had received AI education or training in the past year, compared with 37.6% of those not using it. The census does not say which came first. In accounting, research by Ipsos for Chartered Accountants Worldwide, reported by CA ANZ, found 92% of chartered accountants want AI training and 30% have received any. The study covers chartered accountants worldwide.
AI Uplift Program
Start from how your firm works today
The investigation looks at each workflow and times the tasks inside it, so your first changes go where they help. Your firm keeps the findings, whatever comes next.
3Success as a team
At firm level, the useful methods move from one person into the team's shared workflows. A paralegal's better file note prompt ends up in the hands of the whole team. Work that passes through four people gets reviewed by those four together, and they talk about where AI could help at each step.
People also think about each other when they use AI. Who reads this next? What will they need from it? A partner who knows how a draft letter was put together can review it with more confidence, and the client gets the same standard of letter whoever started it.
Anders Humlum and Emilie Vestergaard studied AI chatbot use among workers in Denmark, and their latest paper found that "take-up rates almost double in workplaces with active employer initiatives." Even without employer initiatives, about 8% of chatbot users had taken on entirely new tasks because of AI, "rising to roughly 17% in workplaces with active employer initiatives."
In the Future of Professionals Report 2026, a Thomson Reuters vendor survey of 1,816 professionals in law, tax, accounting and related fields across 62 countries, 66% of those in firms with a named AI strategy said AI was meeting or exceeding expectations for creating value at work, against 22% where there was no active strategy.
Each person on their own
- AI swapped into existing tasks, one at a time
- Good shortcuts stay with the person who found them
- Change attempted all at once, or put off
- Training picked up wherever people can find it
The team together
- The team reworks one shared workflow at a time
- Methods and prompts the whole team can use
- Small changes, given time to settle
- Time set aside to learn, as a team
4Small changes to start with
Pick one piece of work the team shares.
Client onboarding and the monthly report both pass through several hands. Map who does each step, then talk as a team about where AI could help. A shared prompt library is one way to keep what the team finds.
Train people before asking them to change.
Run a short session on the firm's chosen tools, with the team in the room together. We look at training that sticks in Why the lunch and learn faded.
Put AI on the team meeting agenda.
Spend five minutes on what has worked and what hasn't, including the attempts that didn't pay off straight away. In Microsoft's survey, only 13% of Australian AI users said they are "rewarded for reinvention when results are not immediate."
Let each change settle.
Give a new way of working enough repeated use before judging it, and expect a few lapses along the way.
Write down what the team agrees.
A single page with the agreed tools and who checks what. Over time it can feed into the firm's AI policy, as a draft for the partners to finalise.
5Where we come in
Our investigation measures how long your firm's work takes today, so your team starts from a baseline of its own and can see where the first changes will help. The findings belong to your firm whether or not we go any further together.
If you'd like to talk about where your firm could start, reach out to our team. The first conversation works out whether your firm is the right fit for the AI Uplift Program, and whether we are the right people to help.
Starting the change
Frequently asked questions
What is AI change management?
It is the work of helping a firm's people adopt AI in a way that lasts: training, shared methods, time to practise, and a plan for which workflows change first.
Does a firm have to change everything at once?
No. In Australian research, the productivity boost from computers took around five years of use to work through, and it came alongside skilled people and reorganised work.
How long does a new habit take to form?
In Phillippa Lally's study at University College London, the average time to form a new daily habit was 66 days, with a range of 18 to 254 days. Missing one opportunity did not significantly affect it.
Does AI training make a difference?
In the 2025 Victorian Lawyer Census, 72.3% of lawyers using AI had received AI education or training in the past year, compared with 37.6% of non-users. The census does not show which came first.
What did firms learn from computers?
Productivity Commission research found that computers paid off alongside skilled people, advanced business practices and reorganised work. Early AI research is showing similar associations with employer support, training and clear strategy.
About the author
Julia leads the investigation and is the author of the report presented to your leadership. Her background is in research, data analysis and business strategy.
Primary sources
Australian Bureau of Statistics, Year Book Australia 2002, business use of information technologyabs.gov.auReserve Bank of Australia, Technology Investment and AI: What Are Firms Telling Us?, Bulletin, Fernando, McLoughlin and Ratnayake, 13 November 2025rba.gov.auVictorian Legal Services Board and Commissioner, Generative AI Use in the Legal Profession: Findings from the 2025 Victorian Lawyer Census, 20 April 2026lsbc.vic.gov.auMicrosoft, Aussie workers charge ahead with AI, while leadership falls behind, 2026 Work Trend Index, 15 June 2026 (vendor survey)news.microsoft.comPaul Gretton, Jyothi Gali and Dean Parham, The effects of ICTs and complementary innovations on Australian productivity growth, Productivity Commission, July 2003assets.pc.gov.auErik Brynjolfsson and Lorin Hitt, Beyond Computation: Information Technology, Organizational Transformation and Business Performance, Journal of Economic Perspectives, Fall 2000aeaweb.orgErik Brynjolfsson, Daniel Rock and Chad Syverson, The Productivity J-Curve: How Intangibles Complement General Purpose Technologies, American Economic Journal: Macroeconomics, January 2021aeaweb.orgUniversity College London, How long does it take to form a habit?, on Lally et al., European Journal of Social Psychology, 4 August 2009ucl.ac.ukUniversity of Surrey, Does it really take 66 days to form a habit? We asked the expert, Dr Pippa Lally, 21 January 2026surrey.ac.ukCA ANZ, Empowering CAs to lead with confidence in the age of AI, on Ipsos research for Chartered Accountants Worldwide, 23 February 2026charteredaccountantsanz.comAnders Humlum and Emilie Vestergaard, Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI, NBER working paper 33777, revised March 2026nber.orgThomson Reuters, Future of Professionals Report 2026, June 2026 (vendor survey)thomsonreuters.comGeneral information about how firms organise their AI use. It is not legal, tax or professional advice.