AI versus the human touch: Accountants talk about client demands

Artificial intelligence adoption is a hot topic, but how much technological transformation do clients really want?

Across all industries, the drive to use more AI applications is hard to ignore, but has this changed the demands made by clients? Are more clients expecting accountancy firms to adopt AI solutions, especially if this is perceived to be a way of saving money?

We asked accountants about their working relationship with AI, and whether clients still prefer the human touch.

Clients want the speed of AI but guidance from a professional

Mohammed Sidat, Associate Product Technology Director, Wolters Kluwer Tax & Accounting UK

As technology takes over routine and manual tasks, firms are under growing pressure to evolve their services and deliver deeper advisory value. Our Future Ready Accountant research shows that AI adoption among UK accountants is accelerating, with more than half of professionals now using AI tools regularly – 36% weekly and 24% daily.

Half of the UK-based survey respondents told us that updating their technology has made them more efficient, and 34% report higher client satisfaction as a result. Clients increasingly expect faster turnaround times, clearer insights and digital convenience, but they still rely on the human relationship for reassurance, interpretation and tailored advice.

One key insight from implementing AI is the ongoing challenge of building trust in AI outputs. For example, in automating tax return processes, we consistently see accountants choosing to review and verify AI generated returns, rather than relying on one click automation. This stems from understandable concerns about accuracy and maintaining professional control. It highlights the importance of transparency, auditability and user oversight in AI systems.

Fully automated client advisory or communication without human involvement carries significantly higher risk. AI should support professionals, not replace them, particularly where judgement, nuance and accountability matter most. Ultimately, clients want the speed and efficiency AI brings, but they still expect a trusted professional to guide, validate and interpret the results.

Fully automated client advisory or communication without human involvement carries significant risk. 

Clients don’t want AI for its own sake, but for better outcomes

Luke Thomas, Managing Director, Plus Accounting

Some clients have asked how we’re using AI, particularly those in tech, creative and high-growth sectors who are already using it themselves. Their interest is usually around speed, visibility and insight, such as quicker access to information, clearer reporting or more proactive prompts.

We tend to find that clients don’t necessarily want more AI for its own sake, they want better outcomes, faster responses and clearer decision-making. AI can help enable that, but it still needs to sit within a professionally reviewed framework.

We use AI, but very deliberately and with clear boundaries. We use it to support efficiency and accuracy rather than replace professional judgement. It helps us with data analysis, drafting initial reports, improving internal processes, summarising information and supporting marketing and communications.

It allows our teams to spend less time on repetitive admin and more time focusing on advisory work, strategic conversations and client support, which is where the real value sits. We’re very much at the testing stage of AI in our business and as we all become more confident and familiar with the various platforms we are using, we hope to see even more improvements to our day to day efficiencies.

Many clients are experimenting with AI tools to help them understand financial concepts, draft forecasts or sense-check decisions. That curiosity is a positive thing – it shows engagement with their numbers – but AI can’t fully understand the context of a business, its risk profile or the nuances of tax legislation and compliance.

We often see AI as a starting point for discussion rather than a final answer. Used well, it can lead to better questions, but it shouldn’t replace professional advice.

We see AI as a powerful assistant, not a decision-maker. The balance comes from using AI to improve efficiency, insight and consistency, while ensuring all advice is reviewed, interpreted and delivered by experienced professionals.

The human relationship remains central. Many clients are very clear about retaining the human touch, particularly when it comes to complex decisions, tax planning, business growth, or personal financial matters, clients value reassurance, empathy and experience.

When someone is making a major decision, hiring staff, investing, exiting a business or managing cashflow pressure, they want to speak to someone who understands their situation, not just a tool generating outputs. AI helps our team work smarter, but it’s our people who apply judgement, ask the right questions, challenge assumptions and build long-term trust with clients. The combination of technology supported by human expertise delivers the best outcomes.

AI is a starting point for discussion rather than a final answer.

There’s an opportunity to provide premium service with the human touch

Rebecca Trudgett, Director and Founder, Switchfoot Accounting

Like many accountants, I have been using automation for many years. There are real benefits – AI-wise, I use tools like ChatGPT for research, especially for our impact projects. There are more and more so-called AI features built into accounting software, and if they do the task, especially when they are not client-facing and are safe to use, then I will use them.

But I am in no rush to adopt AI too heavily, and I am happy to adopt it slowly as the technology evolves. Let others test it.

Clients haven’t requested we use AI more. I have noticed clients are using it in the way they use Google, such as looking for tax savings ideas. The problem is that the AI suggests things that are wrong or aren’t tailored to the clients’ circumstances, and then I spend time explaining why the AI is wrong.

AI bookkeeping solutions – which I am sure clients will start to use – are no different to a client doing a bad job reconciling a bank account. I suspect the AI will improve, and with e-invoicing coming, we may see a big improvement in the bookkeeping done by others. But there will always be a role for talented bookkeepers to clear up the mess and sense-check the pickle! There is an opportunity for accountants to support clients and train them to use these tools.

I have been writing an AI policy/decision tree to guide when we use AI and when we avoid it. I am considering how we consistently check the results. In particular, I am concerned about the impact of AI on people and the planet, including energy and water consumption, authoritarianism and wellbeing. And I am concerned about the use of AI, the amount of outsourcing and offshoring in the profession, and the impact of this on the next generation.

There needs to be a balanced, real discussion of the risks, rather than only the rewards to shareholders. I am concerned that AI creates a resiliency threat to a practice, as technology is controlled by so few individuals and corporations.

I run a small, personal and service-led practice, so clients prefer the more tailored approach. For this reason, I have avoided systems that automate without the human touch.

I hear so many people moan about the impersonal service they receive, and there is an opportunity to offer the premium service by offering the human touch. We find that balance by setting up tailored automations. It is essential to be able to review all communications and tweak if necessary. The ability to tailor to individual clients is important.

There needs to be a balanced, real discussion of the risks, rather than only the rewards to shareholders.

Georgia Lewis is a journalist who has worked in Australia, the Middle East and the UK. Over 30 years, Georgia has covered a diverse range of subjects and industries, including business, insurance, technology and logistics..