Artificial Intelligence (AI) is changing the way accounting and tax services are delivered. At our firm, we use carefully selected AI technologies to enhance the quality, efficiency and consistency of the services we provide.

AI enables us to automate repetitive tasks, analyse information more quickly and support our thinking with powerful computational tools. However, it is exactly that, a tool. It does not replace our professional judgement, experience or responsibility to our clients.

Protecting Your Data

The security and confidentiality of our clients’ information is paramount.

Before introducing any AI platform into our business, we undertake appropriate due diligence to understand how it operates, how it processes data, its security controls, and how it complies with relevant data protection requirements.

We only use AI technologies that meet our standards for security, privacy and reliability, and we continually review the platforms we use as technology and regulatory expectations evolve.

Your Data Is Not Used to Train AI Models

A common concern surrounding AI is whether information entered into an AI platform becomes part of the model’s future training.

We have implemented appropriate technical and operational guardrails to ensure that client data is not used to train large language models (LLMs). Where AI platforms process client information on our behalf, we use configurations and services designed to ensure that submitted data remains confidential and is not incorporated into public or shared AI models. This means your information remains your information.

Deterministic AI – When Accuracy Is Critical

Not all AI is used in the same way.

Where a task involves calculations, rule-based analysis or the application of clearly defined logic, we use AI in a deterministic manner. A deterministic process means that, given the same inputs and the same rules, the system will always produce the same result. Examples include applying tax rules, reconciling data, identifying inconsistencies or performing structured calculations.

In these situations, AI is helping us process information more efficiently while producing consistent and repeatable outcomes.

Probabilistic AI – Supporting Professional Judgement

Some tasks are less about calculation and more about interpretation.

Drafting documents, summarising information, identifying themes or suggesting alternative approaches all involve a degree of probability rather than certainty. This is known as probabilistic AI. Rather than calculating a single “correct” answer, probabilistic AI predicts the most likely or helpful response based on the information provided. It can generate excellent first drafts, identify patterns or suggest ideas, but its outputs require professional review and validation.

We use this capability to support our work not to replace our judgement.

The Human Harness

At the heart of our approach to using AI is what we call the Human Harness.

Just as a harness allows a person to direct and safely control a powerful machine, our Human Harness ensures that AI remains under the supervision of experienced professionals at every stage of our work.

Our people decide when AI should be used, what information is appropriate to provide, how its outputs should be interpreted and whether those outputs are suitable for our clients’ circumstances. For the first six months of our year to 31 March 2027 our team stress tested the AI platforms we use so that we understood where our boundaries were.

Every significant AI-assisted output is reviewed by a member of our team who applies professional expertise, commercial understanding and critical thinking before it forms part of the advice or work we deliver.

AI provides speed, scale and computational power. Our people provide context, judgement, accountability and responsibility that makes our use of AI safe, governed and repeatable.

The Human Harness is the combination of governance, expertise and professional oversight that ensures AI enhances our service rather than replacing the trusted advice our clients expect.

Technology with Accountability

We believe AI should make our service better, not less personal.

By combining carefully governed AI technologies with experienced professionals, we are able to deliver work more efficiently while maintaining the high standards of quality, confidentiality and professional judgement that our clients rightly expect.

At Nicholsons technology helps us work smarter; people remain responsible for every outcome.