Building a Confident AI Team

Confidence with AI can look different from one desk to the next. Some people may have already found two or three uses that save them time each week, and they would happily show a colleague how.

Others may know they should be using AI and are holding back. Some are nervous about it, some tried it once, got a strange answer and put it down, and some have Copilot open all day without being sure what it is allowed to see or whether its answers are right.

1How people feel about AI in the workplace

In the Ipsos AI Monitor 2026, 67% of Australian adults said products and services using AI make them nervous, level with Canada for the highest of the 32 countries surveyed. Excitement came in at 36%.

Researchers at the ANU surveyed 3,577 Australian adults in October 2025 for AI Adoption in Australia, a partnership between Google and the ANU. Of those surveyed, 57% agreed that a lack of training or skills limits their use of generative AI.

In Employment Hero's AI Paradox Report 2026, a vendor survey of 1,634 Australian workers, 42% said using AI to complete parts of their job feels like cheating.

Inside a single workplace, the spread can be wide: a Gartner survey of 574 Australian employees in early 2026 found 17% used AI heavily and felt positive about it, and 56% used it little and felt negative about it.

When we talk about confidence with AI, we mean knowing which tools you can use for your work, what they do well, where they fall down, and what is allowed to go into them. It also means being comfortable saying that you use them.

2Why it matters to a firm

Some of that nervousness is about jobs. Randstad's Workmonitor Pulse, a vendor survey run in October 2025, found 32% of Australian workers are worried about AI-driven job loss. The Productivity Commission, drawing on Jobs and Skills Australia's research, reports that for most occupations, AI has a greater likelihood of "augmenting" a task than of "automating" it. We look at what that means for lawyers and accountants in separate posts.

Not knowing how to use the tools can feel like being handed an impressive new program and left to work it out. In the National AI Centre's adoption tracker for December 2025 to February 2026, 19% of small and medium businesses said they don't know how to use AI in their business. Among 105 medium to large firms surveyed by the Reserve Bank in 2025, nearly 40% reported minimal use so far, with firms pointing to a lack of knowledge about the possible applications.

When people work it out on their own, they tend to work it out differently. Employment Hero found 34% of Australian workers use AI tools at work without their employer knowing, and 51% picked up their AI skills from social media. In New Zealand, the same research found 28% using AI without their company knowing. That can mean different tools and different standards for the same piece of client work, which is how shadow AI tends to begin.

Almost two-thirds of Australians in the ANU survey said concerns about how their personal data is used limit their use of AI. In a firm holding client files, people need to know which tools are approved and what client information can go into them.

3One person out in front

A natural place to start can be the person who is already good at it. What we tend to see is that the gains stay in that person's corner of the work. The work around them keeps its old pace, so the bottleneck moves to the next person in the chain. Colleagues may also start sending their AI questions to the one person who knows, which adds to that person's load.

A confident person can also head off in their own direction, with methods and tools the rest of the team can't follow. Building a Champion AI Team covers how that individual skill can become a shared method across the firm.

In a study of customer support agents published in the Quarterly Journal of Economics, Erik Brynjolfsson, Danielle Li and Lindsey Raymond found an AI assistant raised productivity by 14% on average and by 34% for novice and lower-skilled workers.

Bringing a whole team along is hard to do, and the gaps can be hard to see. Randstad found 48% of Australian workers agree they have received training on the AI tools that will reshape their jobs, while 56% of employers say their workforce has been trained. In Microsoft's 2026 Work Trend Index, 28% of the 2,000 Australian AI users surveyed said their organisation is clearly aligned on AI strategy and policies. Among professionals whose firm has a formal AI strategy, Thomson Reuters found 35% say it is not visible in their day-to-day work.

4What a confident team looks like

A confident AI team shares a starting point: which tools are approved, what they do well, where they fall down and what can go into them. The National AI Centre's Guidance for AI Adoption sets out the same idea: making sure the people using or overseeing AI "can understand each system's capabilities, limitations and failure points, and when to intervene".

Knowing where a tool falls down matters as well. In a 2023 Harvard Business School experiment, consultants who used AI on a task it handled poorly were less likely to reach the correct answer than consultants working without it.

People also know who in the team can do what. If the person in accounts has worked out a good way to check a reconciliation with AI, the rest of the team knows to ask her, and her method can become the way the team does it.

From there, each person learns the tools that fit their role. The marketing team may use different AI tools from the research team, and both may use different tools from accounts. Content creation can stay with the people whose job it is. Our post on skills by role goes through it from reception to the partners.

5Where confidence comes from

In the 2025 Victorian Lawyer Census, 72% of lawyers using AI had received AI training in the past year, compared with 38% of those not using it. The census shows the two go together, though it can't show which came first. In total, 70% of the 1,887 lawyers who responded felt they would benefit from AI training or learning materials.

Workplace AI guidelines existed for 42% of the Victorian lawyers surveyed. Of those using AI, 63% worked where guidelines existed, compared with 31% using it without them.

In Denmark, Anders Humlum and Emilie Vestergaard studied about 25,000 workers in 11 occupations exposed to AI, including accounting and legal work. Where the employer had done nothing to encourage it, about 41% had used AI chatbots at work. Where the employer encouraged use, provided a tool and trained people, 93% had used them and 28% used them daily.

Microsoft found organisational factors such as culture, manager support and talent practices were associated with about twice the reported AI impact of individual factors (67% against 32%). When Salesforce and YouGov asked 2,132 Australian and New Zealand knowledge workers what would make them comfortable using AI at work, 45% named access to expert support and 40% easy access to approved tools.

When a team talks regularly about what is working and what went wrong, people have somewhere to ask questions and share what they have learned.

The approved tools, the training and the written guidelines are decisions only a firm can make, and once they are made, people can stop guessing.

6Four ways to build confidence

7Where we come in

Our investigation includes a whole-of-firm survey and conversations across the firm, so you can see how confident your people feel with AI and what they can already do. It also measures how long the firm's work takes today, so the firm starts from a baseline of its own. The findings belong to your firm whether or not we go any further together.

From there, every person completes AI Foundations before cohort workshops grouped by team or skill level, and an AI Champion in each team keeps it going after we leave.

If you'd like to find out how confident your team feels with AI, get in touch. 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 confidence