From experimentation to execution

Why fund industry collaboration is key to realising the benefits of digital innovation

Explore how funds are moving from experimentation to execution with AI, tokenisation and cybersecurity, and why collaboration is key to digital innovation.

6 minute read time

The funds industry has always been complex. Regulation, reporting, governance, investor expectations, cybersecurity and cost pressures already place significant demands on firms. Now, a fast advancing wave of next generation technology is adding another layer - creating new opportunities, but also raising important questions about risk, value and readiness.

These opportunities and challenges were the backdrop for a recent RBS International event, where a panel of experts explored three of the most significant themes currently shaping industry conversations: the potential impact of generative AI and, increasingly, agentic AI; blockchain and tokenisation and cybersecurity. This article brings together the main themes from that discussion, informed by RBS International’s latest research, to explore where the industry is focusing its attention and how firms are beginning to move forward.

The research found that 71% of alternative investment funds see digital transformation as their greatest challenge. At the same time, funds are clear about what they want technology to help them achieve: more automation, stronger compliance and better cybersecurity. Yet only 40% believe their data is good enough to support advanced automation and analytics, while 79% say cybersecurity risks are slowing investment in innovation. The appetite is there but so are the challenges.

Why agentic AI changes the conversation for funds

The starting point in any conversation around AI is the distinction between generative AI and agentic AI in simple terms. Generative AI produces outputs. It can summarise, draft, analyse and respond to prompts. In simple terms, it is a powerful assistant. Agentic AI goes further. It is designed not just to generate an answer, but to take actions towards a goal, following rules, interacting with systems and carrying out tasks on a user’s behalf. As Lewis Lane, RBS International’s Strategy and Innovation Manager put it during the discussion, generative AI is like an intern - somebody you instruct to do work and don’t expect to go beyond the boundaries you set; agentic AI is closer to a junior colleague that can think for itself, operate within a process, bring back information, flag issues and help move work forward.

That distinction matters in the funds world because much of the work is process-heavy, rules-based and often repetitive, but also highly regulated. Think about onboarding, due diligence, policy checks, document review, reporting or exception management. These are precisely the kinds of activities where firms are looking for greater efficiency without lowering standards.

RBS International is already developing agents to undertake some of these tasks. Lewis Lane described work underway around onboarding, where an AI agent is being developed to review applications against procedures, policy and guidance. Crucially, this is not a case of simply letting the machine get on with it and hoping for the best. The process includes training, monitoring, confidence scores and the ability to see the rationale behind what the agent has done. In a regulated setting, that combination of automation and oversight is the point. It allows firms to move at a pace they are comfortable with, build trust gradually and gain confidence before using similar approaches more widely.

Control is the key to successful implementation

Agentic AI may be powerful, but it is not something to be allowed to run unchecked. Andrew Morfill, CEO of M5 Solutions, made the point plainly: large organisations, especially banks, cannot afford to let users build and deploy agents across the IT estate without controls. Access, permissions, data boundaries and monitoring matter just as much here as they do anywhere else in technology. The panel’s message was to start where the use case is clear, the risks are understood and the controls are strong.

That is why the idea of the “human in the loop” remains so important. The panel described a spectrum, from fully human-led processes, through human oversight, to more autonomous workflows. For most firms in funds today, the sensible ground is somewhere in the middle. Agentic AI can help reduce manual effort and improve consistency, but it still needs supervision, clear rules and the ability for people to step in. RBS International’s own recent thinking on AI echoes this approach, suggesting that AI works best when it is integrated thoughtfully into workflows, designed for trust and used to support people rather than replace them.

It’s important to steer the conversation about AI away from an obsession with cutting headcount. Tony Moretta, CEO of Digital Jersey, made a broader point about responsible AI, particularly in a place such as Jersey, where businesses often struggle to fill entry-level roles. In that context, automation is not simply about removing jobs. It can also be about supporting growth, improving productivity, reducing drudgery and helping people move into higher-value work. That is a more constructive way to think about the technology. Agentic AI should be treated like a “synthetic worker”: scoped properly, trained carefully and supervised until it earns trust.

Good data and process foundations are essential

If there was a recurring note of caution, it centred on data or, more accurately, the lack of good enough data. The RBS International research is clear on this point: poor data quality remains one of the sector’s biggest constraints, and many firms know it. Andrew Morfill captured it well: AI amplifies what you already have. If your processes are poor, your documentation weak or your security patchy, AI does not magically fix them, it simply accentuates the problem. The RBS International research makes the same case, warning that without better quality data, emerging technologies will disappoint.

Making sense of blockchain and tokenisation

There is still some nervousness and a lack of understanding around blockchain, tokenisation and digital assets, so the panel set out a simple explanation: blockchain is the infrastructure, tokenisation is the process, and digital assets are the output. It is about creating trusted digital representations of value that can be issued, transferred, tracked and, in some cases, programmed in more efficient ways. For funds, the long-term significance is obvious. Tokenisation has the potential to improve liquidity, reduce friction, widen access and support new operating models.

Lewis Lane pointed to the work NatWest Group is doing around tokenised deposits and digital money, while the research report highlights broader industry expectations that blockchain and tokenisation could improve automation, productivity and even cybersecurity. But the same barriers appear here too: regulatory uncertainty, privacy concerns and the lack of common standards. The solution is for the industry and regulators to find common ground and effective ways to collaborate on how to move forward.

The rise of cybersecurity risks

The report found that more than nine in ten funds expect cybersecurity risks to rise between now and 2030, with 81% identifying increased use of AI by cybercriminals as a major driver. AI is making attacks faster, more convincing and more scalable. Deepfakes, phishing, identity fraud and automated attacks are all becoming more sophisticated.

That means innovation and security cannot be discussed separately. Organisations need to take an active role in strengthening their defences — ensuring their internal policies, controls and data protections are robust, clearly understood and consistently applied across the business. But responsibility does not stop at the organisational level. Individuals also have a critical role to play, by staying informed, understanding their own obligations, and actively building their knowledge through external training, professional bodies and industry associations. As cyber threats continue to evolve, resilience will depend as much on people and skills as it does on technology and controls.

A practical path forward

So, how should funds and financial services businessess adopt and adapt to the technology. Is it a leap of faith or a more considered approach?

The panellists were agreed on the approach. Start with a real problem. Choose a repeatable, high-volume task that consumes time and is prone to human error. Make sure the data behind it is robust. Put guardrails in place. Keep a human in the loop. Measure what improves, then build from there.

The wider message was not that every fund business must become a technology pioneer overnight, but that the direction of travel is now unmistakable. Generative and agentic AI, digital assets and tokenisation, and the growing demands of cybersecurity are converging to reshape how the industry operates.

The opportunity is significant, but so is the responsibility to approach it in a structured and considered way. Firms that take the time to build strong data foundations, audit their processes, put the right controls in place and adopt these technologies with clear purpose will be best placed to move forward with confidence.

If you would like to discuss any of the themes covered in this article further, please reach out to your Relationship Director.

Posted in:

Keywords:

From experimentation to execution

6 minute read time

Latest insights

From experimentation to execution

Agentic AI: What it means for funds and why 2026 is the year to pay attention

Liquidity under pressure: Fund strategy in a higher-rate environment