You went from crypto lawyer to tech founder. What made you realise the solution was not advising clients on compliance, but building the infrastructure to make it automated?
I’ve always been driven to do things better. When ChatGPT was released, it opened my eyes to how legal, risk, and compliance work could change for the better.
I’d been running my crypto-specialist law firm for a few years, advising on high-stakes, billion dollar matters as a sole principal, exhausted but energised from working in a field I love, and desperate for efficiencies.
I tinkered for a while with AI and had a lightbulb moment when I built my first chat application — it helped me do my legal work faster and better, and more affordably for my clients.
My clients thought I was superhuman and to some extent I was.
The app helped amplify my output — both quality and quantity — as a lawyer, but the brain fatigue by end of day also amplified.
I realised that AI needed to shape how humans could serve many rather than have humans take on more work and more 1:1 matters.
And so began the concept of tethering AI agents to domain-expert lawyers and building out a better system which has eventuated in the Crystal agentic Operating System.
You have described Crystal aOS as an ‘agentic operating system’ — can you walk us through what that means in practice for a crypto firm trying to navigate compliance?
In practice, an agentic operating system is the layer that lets a regulated business safely put AI in charge of routine compliance work — without losing human control.
Normally, compliance teams use AI like ChatGPT as a smart assistant.
A human asks a question, reviews the answer, decides what to do next, and moves the process forward step by step.
An agentic system is different. Instead of a human directing every step, the firm defines the process once, and an AI agent runs that process on an ongoing basis — across documents, systems, and tasks.
For example, take annual anti-money laundering (AML) and counter-terrorist financing (CTF) policy updates.
Today:
• -A Compliance Officer uploads the policy into a chat tool.
• -Asks for suggested changes.
• -Manually checks the output.
• -Prepares a summary for approval.
• -Distributes updates and tracks completion of training.
With Crystal aOS, the firm defines that workflow once.
• -An AML/CTF AI agent monitors legal and regulatory changes.
• -When an update is due, the agent drafts changes, prepares a comparison, and packages them for review.
• -Humans approve or reject the outcome.
• -Every step is logged and auditable.
The critical point is control and safety. AI is not trusted to act freely. Crystal aOS puts guardrails in place so:
• -AI agents operate only within approved knowledge and permissions.
• -Outputs are evaluated for consistency and accuracy.
• -Humans remain accountable for decisions.
• -Regulators can see a full audit trail of what happened and why.
So when we say agentic operating system, we mean:
A system that turns compliance processes into supervised, AI-run workflows — reducing manual work while keeping humans, boards, and regulators firmly in control.
What specific compliance pain points are they trying to solve? What is the cost of non-compliance they are trying to avoid?
Legal, risk, and compliance operators haven’t been able to scale themselves — they’re still the bottleneck in fast-growing regulated businesses, even if amplified with AI legal tools. Bottlenecks throttle productivity and business growth. We remove the bottleneck by enabling one human to many agents, and in so doing we allow businesses and innovation to remain on top of compliance and thrive.
Non-compliance can lead to fines and even shutdowns by regulators but the true cost is loss of customer confidence and compliance teams burning out or opting out of the firm. With Crystal aOS, compliance becomes scalable, continuous, visible, and auditable, rather than reactive and understaffed.
That allows firms to maintain a genuine culture of compliance without slowing the business down.
How does Crystal aOS handle the challenge of multi-jurisdiction compliance? If a firm operates across the US, EU, UK, and Asia Pacific, how does your platform manage those conflicting requirements?
Crystal aOS allows law firms to productise their expertise into compliant AI services for clients. Fundamentally, Crystal aOS is regulation/domain-agnostic at the platform level.
Either a crypto firm client or a law firm client can create knowledge sets (e.g. US crypto securities laws), link them to AI agents and deploy the agents into agentic workflows to manage legal, risk and compliance.
If a law firm client has created the knowledge set, agent and workflow, they can licence it to their clients for use in each client’s Crystal aOS environment.
We maintain a global crypto legal and compliance knowledge base that can be accessed by chatting with or using the Crystal master agent, and have a collection of agentic, automated and manual tools to keep the database updated regularly for 195 jurisdictions.
What is your roadmap for expanding from crypto compliance into traditional financial services compliance? How will those two worlds need to coexist?
We’re starting with crypto firms because they have the highest regulatory intensity, fastest change, and strongest demand for scalable compliance. If our system can operate safely there, it can operate in most (if not all) regulated businesses. The natural next adjacency is traditional financial services as those firms start to offer crypto products and services but that also expect a more mature, battle-tested system.
Australia’s ASIC just expanded digital asset relief for stablecoin intermediaries. You recently commented that industry has historically viewed tokens differently than regulators — can you explain that gap and whether ASIC’s approach is closing it?
The gap between industry and Australian Securities and Investments Commission (ASIC) has been longstanding but is closing.
The gap is one where industry is calling for fit-for-purpose regulation of harms specific to crypto without square-peg-round-role regulation-making-by-enforcement-actions against high-signal good actors, and ASIC seeks to use existing laws and existing powers to keep investors and markets safe.
In short for the stablecoin relief, there is a technical view that a stablecoin is not a financial product so does not need relief. More importantly though, there is overwhelming support in industry for stablecoin specific law, like what we’ve seen in the US with their GENIUS Act passed in 2025, because the harms that could occur to consumers and markets from stablecoins are not dealt with by existing laws.
ASIC’s issuance of the relief closes the gap in a sense because the outcome of the relief brings us back to the industry position and draft law has been proposed for stablecoins late in 2025 but continues to be subject to consultation.
Are regulators moving fast enough, or is there still dangerous regulatory arbitrage happening?
Most regulators are moving as fast as they can with the tools they have. Most need larger budgets and better resources to keep pace with innovation and more so for how it is being mis-used.
It’s great to see regulators moving away from regulation-by-enforcement agendas against well-meaning actors and allocating resources to stop bad actors.
For DAOs specifically, what is the biggest compliance challenge they face that traditional entities do not?
A true decentralised autonomous organisation (DAO) is not an entity, and is often described as alegal — neither clearly legal or illegal but just outside reach of the law. This ‘non-entity’ status has been the major challenge for legal recognition and taxation of DAOs that traditional entities don’t have any issue with.
As DAO treasuries custody billions of value, and DAO governance have varying levels of control over treasury spending and protocol safety, governments and regulators have been cautious to ensure the new economic form of capital formation and global governance with global impact has appropriate protections for users and markets.
What needs to happen for institutional adoption to truly accelerate — is it regulatory clarity, infrastructure, or something else?
Leading institutions like BlackRock and J.P. Morgan have been public about how fast and extensively they wish to adopt blockchain technology and digital asset innovation as the new global financial infrastructure. I’m not sure institutional adoption will accelerate en masse. I think those lagging probably lack an internal digital assets champion to lead the change and will fall further behind without a champion or a plan. The sagacious few early movers like BlackRock will soak up market share as laggers leave or liquidate in the next few years.
If you could wave a magic wand and fix one thing about global crypto regulation tomorrow, what would it be?
If I could wave a magic wand, I’d mandate that every government, regulatory body, and courthouse publish laws, guidance, and case judgments to a common repository so that AI applications can easily and accurately search and retrieve from that repository. This should be a unified public good rather than left to the private sector to solve many times over and still in a fragmented sense.
For firms reading The Digital Assets Edge who are sitting on the sidelines waiting for more clarity, what would you tell them?
In the same way that perfection is the enemy of progress, “waiting for more clarity” is the enemy of ongoing relevance. Firms not engaging in digital assets innovation, which is also converging heavily with AI, risk irrelevance of skills and their business model in this next era of agentic commerce powered by programmable digital finance.
There’s more clarity now than last year, and there’s never been more clarity than now. Firms sitting on the sidelines need to get moving yesterday.
← Previous interview
Zumo
Daniel Taylor
.jpg)
Image: Crystal aOS 
