What HMRC’s new AI guidance means for customs professionals

HMRC’s recent guidance on the use of generative AI in tax software is an important development, not only for software developers, but also for customs and compliance professionals who rely on these tools every day.

If you are using AI-enhanced software to assist your customs processes, from data extraction to goods classification and regulatory validation, the key question becomes: 
Can you trust it and can you defend it?
That is ultimately what this guidance is about.

AI Is encouraged, but not unchecked

HMRC recognises the potential of generative AI. It signals that regulators understand innovation is happening and that it can improve compliance processes.

However, the encouragement comes with clear expectations: AI must be transparent, supported by reliable source data, designed with human oversight, secure by design and ethically deployed.

For customs professionals, this should be reassuring.

The message is not “automate everything”.
The message is “automate responsibly”.

No black boxes in a compliance environment

One of the strongest themes in the guidance is transparency and auditability.

If AI is being used, users must:

• Know that it is being used
• Understand its capabilities and limitations
• Be able to identify the source data
• Be aware of the risk of inaccuracies or hallucinations

In customs operations where compliance is key, this matters enormously.
If a classification decision, origin interpretation or valuation approach is ever challenged during an audit, you need to understand how that conclusion was reached. 

An answer generated by a system without traceability creates exposure. AI must therefore be explainable enough to support audit defense. Technology should not weaken your position. It should strengthen it.

Reliable source data is non-negotiable

HMRC is clear: AI-enhanced software must rely on authoritative sources such as legislation, official publications and established case law.

For customs professionals, this is fundamental.
Customs regulations vary by region and are constantly evolving. Any system supporting your work must reflect those changes promptly. Otherwise, the risk does not sit with the software provider, it sits with the declarant and the business submitting the declaration. 

That is why continuous monitoring and timely updates are not technical features. They are risk controls.

Human oversight remains central

Perhaps the most important aspect of the guidance is the emphasis on human oversight.

AI should support, not replace, professional judgement.

In practice, that means:

• Complex scenarios should be flagged clearly
• Nuanced rules should trigger review prompts
• Users must be able to correct outputs easily
• Responsibility for accuracy remains explicit

For customs professionals, this reinforces something you already know: accountability does not disappear because a system is involved.
Well-designed AI should reduce repetitive workload, highlight potential issues and improve consistency but it should never remove your control.

Data security and privacy matter more than ever

Customs processes involve highly sensitive commercial data. The guidance emphasises secure software development, compliance with GDPR and privacy by design.

This is not theoretical.
When AI tools process declarations, logistic documents or customer data, security must be embedded at every stage. Trust in AI is inseparable from trust in how your data is handled.

What this means in practice

For customs professionals using AI-supported systems, HMRC’s guidance offers clarity:

• You should expect transparency.
• You should expect traceability.
• You should expect continuous updates aligned with legislation.
• You should expect the ability to review and override outputs.
• You should expect strong data protection.

If any of these are missing, the risk remains yours.

At Customaite, we see this guidance as reinforcing a principle we strongly believe in: AI in customs is not about replacing expertise. It is about strengthening it.

When repetitive logistic document extraction, classification checks or rule cross-referencing can be handled efficiently by a system, professionals gain time to focus on complex cases, exceptions and strategic advisory work. 

But control must remain where it belongs, with you.

Responsible AI is in everyone’s interest

AI is increasingly part of the customs landscape. HMRC’s publication makes it clear that innovation is welcome, provided it is responsible and transparent.

For customs professionals, this is good news.
It means the regulatory direction supports tools that:

• Improve consistency
• Reduce manual workload
• Enhance compliance
• Maintain accountability

AI should not introduce uncertainty into your processes.
It should reduce it.

That requires technology designed with traceability, up-to-date regulatory grounding and meaningful human oversight built in from the start.

At Customaite, we work closely with customs professionals to ensure AI strengthens, rather than complicates, their compliance processes.

If you would like to explore how responsible, transparent AI can support your operations while keeping you firmly in control, we would be pleased to speak with you.

Get in touch with us at Customaite to discuss how we can support your team.