AutoRek CEO: AI Speed Is No Substitute for Financial Controls

Financial services technology buyers have spent years wrestling with applications and data spread across fast-moving businesses. The integrations that sit on the edges of these systems are costly to build, troublesome to maintain, and carry real operational risks around data security and resilience. That complexity spans different technology stacks, varying levels of interoperability, and a heavy reliance on legacy systems of record.

The appeal of AI is straightforward: a way to work across those estates with minimal change to the underlying architecture, creating new insight and unlocking value. It is tempting to treat AI as a universal solvent that dissolves complexity without touching the systems beneath. Chris Livesey, chief executive of financial controls software firm AutoRek, argues the real test is not how quickly AI gets adopted. The test is whether it protects the controls the sector depends on.

Some buyers are asking whether they can remove applications altogether, replacing them with self-built agentic services that meet the same need across their domain. That would free them from licence costs and let them work at the speed of their business rather than the speed of a supplier. Software vendors, in turn, see disaggregating their own applications into agentic services as a way to keep offering their expertise and to protect their businesses from customer self-build.

Large enterprise software stacks will almost certainly be broken up into capability platforms that a hybrid workforce of people and AI agents can operate. Those capabilities will sit alongside a wider ecosystem, orchestrated centrally where possible and underpinned by an operating-model control plane spanning data quality, governance and interoperability. The competitive ground will shift from the edges of applications to the ownership of agentic orchestration and decision-making.

That control plane is where agents and humans meet, and how it works in practice is still unclear. It is also where buyers and suppliers will need to converge. If they do not, disconnected strategies risk fragmenting approaches, standards and methods, and increasing fragility in the underlying business services. That is the opposite of the enduring value financial services firms are trying to give their own customers. In financial control, this matters more, because the operating model itself is the product.

In financial controls and regulatory compliance, the work starts with the control and moves outward. The need for provable results has not changed. What has changed is the range of technologies available to deliver them. There is durable value in using AI to shorten implementation, improve configurability and embed intelligence into financial control. The point is to move fast where AI genuinely strengthens those outcomes.

Much of the current AI narrative implies the opposite, treating controls as secondary to the technology and suggesting that regulation will have to adapt to AI rather than the other way around. How regulators and the audit profession respond to an AI world remains to be seen. If their goals stay centred on protecting people and society, that adaptation is unlikely to make easy room for new risk.

AI is an accelerator, not a substitute. The question is not whether it will reshape the operating model, because it will. The question is whether it does so in a way that enhances the duty of care financial institutions owe their customers. Applied well, AI strengthens trust, accuracy and confidence. That is the destination, Livesey said.