01/09/2026
Interview with Christian Crespo, CEO and founder of icloudCompliance
"We believe that compliance in Spain is at a critical juncture following the reform of the Criminal Code, and this discussion needs to take place as a community, not with everyone working in isolation in their own offices."
With over eight years’ experience in the Spanish GRC market, icloudCompliance has established itself as one of the leading technology platforms for comprehensive regulatory compliance management. Its approach focuses on centralisation, automation and continuous updating to help organisations turn compliance into a genuine competitive advantage.
As part of the Compliance 360º Andalucía: People, Technology and Critical Decisions event, to be held on 22 October in Seville, Christian Crespo, CEO and founder of icloudCompliance, will take part in a panel discussion analysing the impact of artificial intelligence on the compliance function. We spoke to him about the evolution of technology, the opportunities opened up by AI and the role that professionals will continue to play in an increasingly digitalised environment.
1. How is AI changing the compliance function?
It is changing the very nature of the role. For years, the compliance officer has been an ‘evidence hunter’: tracking down documents, requesting reports, preparing for committee meetings. When we analysed how our more than 130 clients actually use the platforms, we discovered something revealing: most modules in any compliance software are hardly ever used on a day-to-day basis. The problem was never a lack of tools, but rather that the tools did nothing for the professional.
AI turns that equation on its head. At iCloudCompliance, we’ve summed it up in three verbs: compliance AI should tell you what to do, do it for you, and protect you legally. That’s the standard we’ve built with Brújula, our intelligence layer. The Compliance Officer is no longer merely an administrative compliance officer but becomes what they always should have been: a risk strategist.
2. To what extent is AI replacing operational tasks, and where is the limit beyond which the human factor remains irreplaceable?
Today, AI already reliably handles all operational tasks: gathering and classifying evidence, monitoring regulatory deadlines, drafting committee reports, and cross-checking a transaction against hundreds of controls. We’re talking about tasks that used to take up 60–70 per cent of a compliance officer’s working day.
The limit is very clear, and it is a legal limit, not just a technical one: responsibility cannot be delegated to an algorithm. No AI is going to sit before a judge or assume criminal liability for a decision. Ethical judgement, understanding the organisational context, and the final decision when faced with a dilemma: that is, and will remain, a human responsibility. Our philosophy is that AI prepares the decision; the professional makes it and signs off on it.
3. What role should technology play in critical decision-making?
That of a co-pilot, never that of a pilot. In critical decisions — such as disciplining an employee following an internal investigation, terminating a relationship with a supplier, or reporting an incident to the regulator — technology must do three things: provide all relevant data, present scenarios along with their consequences, and ensure the entire process is properly documented.
This is not a matter of opinion: it is exactly what the European AI Act requires for high-risk systems – effective human oversight. We implement this by design: on our platform, there is no critical decision that AI executes without human validation. And that architecture, paradoxically, is what best protects the professional: if they ever have to explain their decision, they have the complete record of how and on the basis of what information it was made.
4. In day-to-day practice, which compliance process saves the most time today when it is automated and integrated with AI?
Without a doubt, evidence management and the preparation of reports for committees and boards of directors. It is the great time-sink of compliance: weeks spent chasing documents through the post, only to then manually compile a report that the committee reviews in twenty minutes.
With integrated AI, that cycle goes from weeks to hours. A real-life example: a management review report for an ISO system — which traditionally requires consolidating indicators, audits, incidents and targets for an entire financial year — is now generated using the information already available on the platform. And the second major time-saving comes from the whistleblowing channel: Law 2/2023 imposes strict deadlines, and AI ensures that no case is left unresolved. Our clients tell us that this is the difference between sleeping soundly and not.
5. When relying on AI to assess risks, how do you ensure the traceability and legal explainability of its decisions when faced with an auditor or a judge?
This is, for me, the most important question of the day, and where serious AI is distinguished from hype. Our internal rule is categorical: if you cannot explain to an auditor or a judge how the AI reached a conclusion, you should not be using it in compliance.
In practice, this means four things that we have built in by design: every AI recommendation cites its source — the regulation, the control, the specific evidence — it is never a black box; everything is recorded in an immutable audit log with date, version and user details; confidential client data is anonymised before any external processing, using an architecture we have developed specifically for the sector; and the final decision always bears a human signature. The result is that the file generated by the AI is more legally defensible than the one produced manually, not less so.
6. What is the biggest challenge when digitising a compliance programme?
The biggest challenge is not technological: it is methodological and cultural. We always tell our clients: if you digitise chaos, you get chaos even faster. Many organisations come to us with their compliance programme scattered across Excel spreadsheets, emails and shared folders, without a coherent risk and control framework underpinning it. Simply slapping software on top of that solves nothing.
That’s why at iCloudCompliance we don’t just sell a platform: we provide support through our own methodology and, when the client needs it, through our Governance as a Service model, with dedicated engineers working within the client’s operations. First we organise the model, then we digitise it, and only then does AI amplify the results. That is the correct order, and skipping it is the number one cause of failed digitisation projects in compliance.
7. Given the framework of the AI Act and the heavy European regulatory burden, how does technology help to comply with regulations without simultaneously becoming a new legal risk?
It is the great paradox of our time: the only realistic way to absorb the European regulatory burden — the AI Act, CSRD, NIS2, Law 2/2023, data protection — is through technology, but poorly governed technology is in itself a regulatory risk. The answer is to apply compliance to the AI itself.
We lead by example: we have classified our own AI systems in accordance with the AI Act, we document how they operate, we ensure human oversight and, above all, we have designed the architecture so that confidential client data never leaves their perimeter without being anonymised. When an AI provider for compliance cannot explain how it governs its own AI, there’s your answer. Technology that helps ensure compliance is technology that is built to comply from the outset.
8. Why have you decided to support Compliance 360º Andalucía and what do you expect from this conference?
Because we believe that compliance in Spain is at its most critical juncture since the reform of the Criminal Code, and that conversation needs to take place within the community, not with everyone working in isolation in their own offices. The World Compliance Association has been building that community for years and we wanted to demonstrate our full commitment by becoming Gold sponsors.
Above all, we hope this event will foster an honest conversation about AI: separating what is real from what is just marketing. We’re here to showcase specific case studies — how real compliance officers are working with real AI today — and to listen. I invite any professional who is curious, or sceptical – and scepticism is healthy – to come and speak to us in Seville. Bring us your most challenging process and we’ll show you live what AI would do with it.
9. A one-liner for the debate: Will AI replace the compliance officer, or will the professional who uses AI replace the one who doesn’t?
“No AI will go to prison for you: the compliance officer is irreplaceable. But those who use AI will do the work of ten, and the market is already choosing which of the two it prefers.”
