Executive summary
AI skills cover understanding what a tool is doing, using it well in your own work, and knowing when a person needs to check what it produces. This piece looks at what those skills are, why they matter in every role from the front desk to the partners, what people and firms are doing about them at the moment, and how a firm can work out what each role needs.
Learning to use AI at work tends to happen one desk at a time. One person works out how to get a usable first draft from a few lines of notes. Another learns which questions a tool handles well and which it gets wrong. Each of them becomes more capable with practice, and the firm gains from that effort.
In a study by Deloitte Access Economics for RMIT Online of more than 2,000 Australian workers, 84% had used at least one AI tool, and 7% reached what the researchers class as advanced AI literacy. More than half were beginners.
1Understanding and using AI
Treasury's 2026 Intergenerational Report says the skills AI asks for "may range from foundational AI literacy to occupation-specific capabilities and specialist technical expertise", with critical thinking, communication and change management in demand alongside them.
In a firm, foundational literacy means understanding how the tool produces its outputs, what happens to the information you put into it, and the kinds of mistakes it can make. Occupation-specific capability means using it well in your own work, whether that is a letter of advice, a reconciliation or the notes from a client meeting.
The Law Society of NSW sets out the solicitor's version in its guide to responsible use of AI. It asks firms to make sure practitioners have the skills to "appropriately prompt the generative AI tool", to "evaluate the common limitations of the tool", and to supervise its outputs and their accuracy.
The Thomson Reuters Future of Professionals Report 2026, a vendor survey of 1,816 professionals in law, tax, audit and accounting, says: "Most AI training programs teach what the tools do. They rarely teach the harder and more valuable skills: knowing when to trust AI output, when to verify it, and when to override it."
A firm relies on that last skill when it says there is a human in the loop: the person reading the work knows what the tool can get wrong.
When the Victorian Legal Services Board + Commissioner asked 1,887 lawyers in its 2025 Victorian Lawyer Census what AI training would help them, interest was strongest in how AI can enhance workflow and efficiency (74%), followed closely by the safe and ethical use of AI (70%).
2Demand for AI skills
Treasury notes that "some of the most highly exposed occupations include accountants, receptionists and advertising and marketing professionals." For other tasks, it says, "human judgement, coordination, accountability, tacit knowledge and relationships will continue to be integral."
In PwC's Hopes and Fears survey of 1,225 Australian workers, a vendor survey, 50% of non-managers felt they had the resources they need for learning and development, against 75% of senior executives.
In the Victorian census, 70% of lawyers felt they would benefit from AI training or learning materials, and 18% felt they would not. Across the wider population, the ANU's AI Adoption in Australia study, run in partnership with Google, found 57.0% of 3,577 Australian adults agreed a lack of training or skills limited their use of generative AI. The government's National AI Plan reports that the AI microcredential courses at the Institute of Applied Technology have attracted more than 150,000 enrolments.
In the ABS's Characteristics of Australian Business survey for the 2024-25 financial year, 35% of Australian businesses reported a shortage of skills of any kind, and 26% of those had invested in training or retraining the people they already employed.
In a liaison survey of 105 medium and large firms, the Reserve Bank found adoption "has been shallow to date, with nearly 40 per cent indicating minimal use so far." Firms told the bank they expected AI to help lift productivity if they could complement it with skilled people, changes to workflows and culture, and new ways of working.
3Where firms are at the moment
What we tend to see is AI in use across a firm, by people at different stages with it.
In professional, scientific and technical services, the ABS industry group that includes law and accounting, 24% of businesses used AI in 2024-25. Among Australian accounting and bookkeeping practices, Xero's Modern Practice Playbook, a vendor survey of 524 senior accountants and bookkeepers, found 85% of practices had tried AI and 52% used it day to day. In the same survey, 39% of practices planned to put the capacity AI frees up into training and skills.
A Gartner survey of 574 Australian employees found that, of those expected to use AI in their role, 85% were given enterprise AI tools and 86% also used personal AI tools to work more efficiently or take on more complex tasks.
In Denmark, Anders Humlum and Emilie Vestergaard asked about 25,000 workers, including accountants and legal professionals, what their employer had put in place for AI. The share who had used AI chatbots at work ranged from about 41%, where the employer had not encouraged AI, provided a tool or offered training, to 93% where it had done all three. The study shows an association and cannot say which caused which. Each of those three is a decision for the firm, and the person who owns AI is well placed to make it.
AI Uplift Program
See where your firm's AI skills sit
The investigation surveys the whole firm about how people use AI today and measures how long your everyday work takes. You keep the findings either way.
4Skills by role
The Law Society of NSW guide describes AI as needing "constant human oversight, at all levels of your business." An example role map might look like this:
| Role | Understands | Uses AI for | Checks |
|---|---|---|---|
| Reception and front desk | Which tools can hold client details | Replies to routine enquiries, message summaries, diary notes | Names, dates and what the client will read |
| Accounts and administration | What the tool does with figures it is given | Invoice queries, reconciliation drafts, formatting | Figures against their source where they will be relied on |
| Graduates, paralegals and junior accountants | How the tool produces its outputs, and that it can invent a case or a number | First drafts, research summaries, file notes | Citations, facts and figures before the work goes up |
| Senior fee earners | The firm's position on AI for client work | Reviewing drafts, structuring advice | Whether the work is right for this client now |
| Partners and practice managers | Where AI is used across the firm | Planning, reporting, the firm's own documents | That each role knows its checkpoints |
Reception and front desk
- Understands
- Which tools can hold client details
- Uses AI for
- Replies to routine enquiries, message summaries, diary notes
- Checks
- Names, dates and what the client will read
Accounts and administration
- Understands
- What the tool does with figures it is given
- Uses AI for
- Invoice queries, reconciliation drafts, formatting
- Checks
- Figures against their source where they will be relied on
Graduates, paralegals and junior accountants
- Understands
- How the tool produces its outputs, and that it can invent a case or a number
- Uses AI for
- First drafts, research summaries, file notes
- Checks
- Citations, facts and figures before the work goes up
Senior fee earners
- Understands
- The firm's position on AI for client work
- Uses AI for
- Reviewing drafts, structuring advice
- Checks
- Whether the work is right for this client now
Partners and practice managers
- Understands
- Where AI is used across the firm
- Uses AI for
- Planning, reporting, the firm's own documents
- Checks
- That each role knows its checkpoints
The map gives each role a clearer idea of what it needs to know and what it is responsible for checking.
Skills built desk by desk
- Each person learns at their own pace and in their own way
- Methods that work stay with the person who found them
- Access to learning follows seniority
- Checking depends on what each person knows about the tool
Skills built across the firm
- Each role has a clear picture of what it needs
- The firm provides the tools, encouragement and learning
- Every role has access to learning
- Each person knows what the tool can get wrong in their work
5Working out what each role needs
List every role in the firm.
Reception, accounts and practice support go on the list alongside the fee earners.
Ask each role what they use AI for now.
A few questions in a staff survey will do, along with what they would like to learn.
Fill in the map.
For each role, note what it needs to understand, what it uses AI for and what it checks.
Decide what the firm provides.
The tools for each role, encouragement to use them, and training that lasts.
Ask again later.
Use the same questions and compare the answers.
6Where we come in
Our investigation includes a survey of the whole firm, so you can see what each role already does with AI and where people would like more support. It also measures how long your everyday work takes today, so you start from a baseline of your own. The findings belong to your firm whether or not we go any further together.
From there, AI Foundations gives the whole firm the same starting point, and the cohort workshops that follow are grouped by team or skill level.
If you would like to see where AI skills sit across your firm, 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.
Building AI skills role by role
Frequently asked questions
What are AI skills?
Treasury's 2026 Intergenerational Report says they may range from foundational AI literacy to occupation-specific capabilities and specialist technical expertise. In a firm, that means understanding what a tool is doing, using it well in your own role, and knowing when to check what it produces.
What is the difference between AI skills and AI training?
AI skills cover understanding what a tool is doing, using it well in your role and knowing when to check its output. AI training is one way a firm builds them, focused on how to use the tools and get better at them.
Do people want to build AI skills?
In the 2025 Victorian Lawyer Census, 70% of lawyers felt they would benefit from AI training or learning materials. In an ANU study of 3,577 Australian adults, 57.0% agreed a lack of training or skills limited their use of generative AI.
Does every role in a firm need AI skills?
Treasury lists accountants and receptionists among the most highly exposed occupations and expects human judgement to stay integral to other tasks. The Law Society of NSW describes AI as needing human oversight at all levels of a business.
How many Australian businesses are short of skills?
In the ABS Characteristics of Australian Business survey for 2024-25, 35% of businesses reported a shortage of skills of any kind, and 26% of those invested in training or retraining the people they already employed.
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
The Treasury, 2026 Intergenerational Report, 21 September 2026treasury.gov.auThe Law Society of New South Wales, A Solicitor's Guide to Responsible Use of Artificial Intelligence, updated January 2026lawsociety.com.auDeloitte Access Economics for RMIT Online, Beyond Prompting: Measuring the Generational AI Gap, 25 March 2026deloitte.comThomson Reuters Institute, Future of Professionals Report 2026, 22 June 2026thomsonreuters.comVictorian Legal Services Board + Commissioner, Generative AI Use in the Legal Profession: Findings from the 2025 Victorian Lawyer Census, 20 April 2026lsbc.vic.gov.auPwC Australia, Global Workforce Hopes and Fears Survey 2025: Australian insights, 28 January 2026pwc.com.auHerrington and Biddle, AI Adoption in Australia, ANU School of Cybernetics in partnership with Google, 9 July 2026cybernetics.anu.edu.auDepartment of Industry, Science and Resources, National AI Plan, 2 December 2025industry.gov.auAustralian Bureau of Statistics, Characteristics of Australian Business, 2024-25 financial year, 25 June 2026abs.gov.auFernando, McLoughlin and Ratnayake, Technology Investment and AI: What Are Firms Telling Us?, RBA Bulletin, 13 November 2025rba.gov.auXero, Modern Practice Playbook, media release, 29 September 2026xero.comGartner, Gartner HR Survey Finds AI is Boosting Productivity but Widening Workforce Divisions in Australia, 2 June 2026gartner.comHumlum and Vestergaard, Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI, NBER working paper 33777, revised March 2026nber.orgGeneral information about how firms organise their AI use. It is not legal, tax or professional advice.