Our Approach to AI
AI in compliance is changing fast. Here's where we stand.
Every compliance team is being asked the same question right now: what are you doing with AI? It's a fair question. But in shareholding disclosure, where the wrong answer to a regulator can mean fines, reputational damage, and personal liability, it deserves a more considered answer than most vendors are giving. Here's ours.
The reality of AI in compliance
AI can do a lot in this space. It can surface what matters faster, cut hours of manual investigation, and keep pace with regulatory change at a scale no team can match alone.
But it can't read a regulation the way a lawyer does. It can't account for the context your team carries. And when something goes wrong - and in compliance, sometimes it does - it can't stand in front of a regulator and explain the decision.
Regulatory accountability is non-delegable. Someone is always responsible when a filing is wrong. That's not a philosophical position - it's a legal one.
That's the reality we're building around.
What we're doing
We've been using AI in our platform and across our business for some time now - carefully, deliberately, and with real results. Here's what that looks like in practice.
AI for enhanced alert review
AI-enhanced workflows help clients prioritise what matters. By surfacing anomalies, identifying likely false positives, and reconciling disparate data sources, we reduce time spent on manual checking - without removing the human expert from the process.
AI for filing automation
Many jurisdictions now require direct portal submissions, each with their own formats, validation rules, and quirks. We've built systems that streamline filings that used to take hours into a few clicks - while keeping every submission fully traceable and defensible.
AI for rule management
We're exploring AI to assist in regulatory change monitoring and internal ruleset mapping - but not to replace legal experts. Parsing a regulation into an enforceable rule is an act of interpretation, not formatting. We're building AI tools to make our expert team faster, not to replace their expertise.
What we're not doing. And why.
A measured approach for new technology
The idea of "lights-off compliance" - where alerts are detected, validated, filed, and closed without human involvement - is appealing. But realising that vision responsibly requires what most automation vendors overlook: strong governance, intelligent design, and deep compliance expertise at every layer. The barriers aren't just technical.
We're not claiming compliance can run without people
Legal liability still rests with humans. Filing errors can damage market reputation. Auditability and explainability are regulatory requirements, not options. Some vendors claim otherwise. We don't.
We're not using AI to make compliance decisions
AI will surface, summarise, and prioritise. It will not determine whether a threshold has been breached, whether a filing is required, or what position to take on an ambiguous rule. Those calls belong to your team.
We're not training models on your data
Your portfolio data, filing history, and internal configurations are not used to train AI models - ours or anyone else's. Your data stays yours, governed by the same security standards it always has been.
We're not reacting to hype
We only apply AI where it adds measurable value and where it can be done with legal defensibility and smart fail-safes. We're building towards autonomous workflows - but only when it's safe, defensible, and transparent.
Our thinking, in full
We've put our complete position on AI in shareholding disclosure into one place - what we believe, what we're building, and why we think accountability matters.
What our clients ask us and what it means for them
In practical terms, here's what AI changes for you - and what stays exactly the same.
Will this affect the accuracy of my compliance checks?
No. AI features sit alongside the core rules engine, not inside it. Your threshold calculations, disclosure obligations, and filing outputs are generated by the same rules-based system they always have been - maintained by our in-house legal and regulatory experts.
How does AI actually save my team time?
Alert triage that used to mean hours of manual checking, and filings that used to take hours, now take minutes. AI does the heavy lifting - surfacing anomalies, mapping data to filing templates, and handling portal submissions. Your team makes the final call.
Is my data safe?
Yes. All AI features operate within our existing ISO 27001 and SOC 2 security framework. Your portfolio data, filing history, and configurations are not used to train AI models - ours or any third party's. No new data sharing arrangements are created without your knowledge.
How do I know what's been produced by AI versus a human expert?
Every AI-assisted step is clearly labelled in the platform and fully logged. Every submission remains traceable and defensible for regulatory purposes. You will never be left guessing what came from a model and what came from a human.
Want to talk through what this means for your team?
Whether you're navigating an AI directive from above or evaluating what responsible automation looks like in practice, we're happy to have that conversation.