Fear, friction, and false starts: The human barriers to AI

The initial obstacles to AI adoption have little to do with technology. “Ignorance,” as Niels Footman of Dixon Humphreys, AI accelerator consultancy, put it bluntly, remains one of the main barriers, alongside fear of change, uncertainty, and misinformation. But these can be overcome not through avoidance, but through action.

“The way to counter this,” Niels explained, “is not to ignore AI, but to understand its huge opportunities and mitigate its risks.” He emphasised that firms which immerse their people in AI, even in small, low-stakes use cases, quickly reduce fear and build confidence. “If you want to figure out how AI can work best for you and your business, the best way to learn is through trying things out yourself. Just a single successful use case can transform how you view AI.”

Building AI capability starts with purpose, not platforms

One of the most powerful takeaways from the discussion was the idea that AI capability is not just about infrastructure, it’s about talent. At Dixon Humphreys, AI initiatives begin not with a two-year strategy handed down from the C-suite, but with a “people-first” model focused on practical application.

“Start with a purpose,” Niels advised. “Don’t look at all the tools and panic. Start with something in your daily work, even planning a holiday with AI can be eye-opening.”

This purposeful experimentation laid the groundwork for success at Hurst Accountants, where Partner Jo Gibson explored a firm-wide AI education programme. Every employee – from partners to PAs, completed the initial AI accelerator programme to build shared knowledge and understanding. From there, they launched an internal AI operations team, made up of volunteers across departments, who were given dedicated time to “make and break” AI tools through iterative experimentation.

Crucially, Jo urged firms not to underestimate where innovation might come from. “Some of our most surprising contributors weren’t senior leaders or tech people, they were curious staff who started playing with prompts and building bots.”

Security and Trust: why freedom to experiment is enabled by governance

While enthusiasm is growing, firms remain rightly cautious about data protection and AI hallucinations. Larking Gowen, for instance, initially trialled ChatGPT but later transitioned to Microsoft Copilot after assessing the relative security risks.

As Giles Kerkham, Partner and Head of Not for Profit at Larking Gowen explained, “Copilot provided us with traceable outputs and assurance that data wouldn’t be used for training models. It wasn’t just about features, it was about confidence. With those software features and a clear policy framework, we can encourage our people to experiment and use novel solutions.”

They also ran internal surveys to understand usage trends. Surprisingly, the most common reason for not using AI tools was, “I just didn’t think of it.” Culture matters. Seeing someone next to you using Copilot, or even an onscreen Bot, may become the nudge needed to try it yourself.

Beyond efficiency: How AI is enhancing audit quality

While much of the early AI conversation focused on time savings, its potential for improving audit quality is even more exciting. Buzzacott shared a compelling example through their deployment of MindBridge, an AI-powered audit tool that uses machine learning to risk-score every journal entry in a general ledger.

“As client data grew more complex, traditional Excel-based methods simply couldn’t cope,” one manager explained. “With MindBridge, we’re not just auditing the ‘known knowns’ – we’re uncovering the ‘unknown unknowns.” Shared Ben Rennie, Audit Assistant Manager at Buzzacott.

By benchmarking transactions against control points (e.g., weekend postings or unusual rounding) in addition to machine learning, auditors gained deeper insight and consistency. Implementation wasn’t seamless – it required one-to-one manager meetings, RI committee approvals, and targeted training. But the payoff is a clearer, more intelligent audit trail. Crucially, the tool enhances rather than replaces professional judgement. Auditors still interpret the outputs and make final calls, but with a broader, more consistent base of evidence.

Practical innovation: From South Africa’s back office to strategic resource planning

AI’s influence is not limited to audit or compliance. In South Africa, Suresh Naidoo, CEO at Accensis introduced Beeye, a resource planning tool that unified disjointed scheduling across tax, audit, and accounting teams. Their experience highlights a different, but critical, challenge: project management.

“We assumed younger staff would embrace new tools easily,” one executive reflected. “They didn’t.” The lack of experience in staff engagement, structure, and ownership led to delays. Once a committed project lead was appointed, however, adoption surged.

Their story underscores the need for thoughtful change management and the fallacy of assuming tech proficiency equates to adoption.

Lessons for the profession: What leading firms are doing right

From the diverse, but overlapping journeys, we discussed, six key lessons stood out:

1. AI will not replace accountants, but accountants using AI will outperform
   Firms should view AI not as a threat but as an accelerator, one that enables professionals to shift from low-value tasks to high-value insights.
2. Start small, but start now
   Whether it’s onboarding automation, invoice matching, or meeting summarisation, practical use cases help build momentum and confidence.
3. Be intentional about security
   Balancing innovation with compliance is essential. Firms like Larking Gowen demonstrate that security-conscious adoption is possible.
4. Training must be tailored
   From junior auditors to managing partners, different roles require different learning paths. Face-to-face sessions, lunch-and-learns, and internal champions all have a role to play.
5. Create a culture of sharing
   Some of the most impactful learning comes not from formal training, but from peers showing peers what’s possible.
6. Treat AI like a colleague
   As Jo from Hurst put it: “Delegate tasks to AI like you would to a teammate, give it context, clarity, and direction.”
7. Share and shape policies together

Many firms are still early in their journey, with a majority reporting they’ve yet to formalise policies. One roundtable participant emphasised the value in sharing AI usage policies across the profession to prevent unintentional misuse.

When discussing risks, the biggest was seen as staff using AI tools and sharing outputs with clients without proper review. A clear AI policy should therefore not only set out what’s not allowed but also encourage staff to seek sign-off before relying on AI-generated content. As one leader put it, “The fear is not just what AI might say, it’s that someone uses it without oversight and the firm becomes liable.”

Some participants recommended making use of community resources, like shared policy templates and AI discussion forums, as a helpful starting point. For firms just beginning this journey, peer support and policy sharing can make a significant difference.

Final Thoughts: A Profession in Motion

This isn’t just about automation, it’s about transformation. AI is helping accountancy firms move from reactive compliance to proactive, strategic insight. But that transformation only happens when people, not just systems, are brought along.

At PrimeGlobal, we see AI not as an existential threat but as a historic opportunity. One that will reward firms who are willing to learn, share, and lead with purpose.

As one speaker rightly said: “There’s never been an easier technology to try, just open it and talk to it. You might be surprised by what it can do.”