Agents are getting all the press, and rightly so. They promise something genuinely different. End-to-end execution at a lower cost than the human alternative. That’s a compelling proposition.
In practice, fully autonomous use is rare. UK financial services isn’t ready for it. It’s not clear it ever will be.
There are a few reasons:
One. Autonomy is hard.
To run autonomously, a system needs clear process boundaries, reliable data and consistent escalation paths. Without those, autonomy introduces risk rather than efficiency.
Two. Structured workflow is missing.
Coming back to the earlier point, without structured workflow firms lack visibility, control and predictability. Autonomy on top of that just amplifies the problem.
Three. The maturity model is oversimplified.
There’s a tendency to frame AI adoption as a linear progression. Master Copilot, move to agents, then full autonomy. In reality, as capability increases, control and governance needs increase too. Adoption gets harder, not easier. And not every solution needs to be fully autonomous. A human-orchestrated agent can still deliver meaningful value.
Where firms should focus
Simplify it. Automation defines the process. AI improves decisions. Orchestration connects the two. Autonomy sits at the far end. It isn’t the starting point.
This leads back to the uncomfortable view. Most organisations don’t need full autonomy. They need assisted workflows that remove mundane activity and free up capacity. That allows teams to focus on higher value customer interactions and increase output without hiring.
These use cases can be deployed safely. They give firms space to test what works and what doesn’t. That learning can then be applied to more complex scenarios. This is where value is being realised today.
There’s also the human factor. Organisations can only absorb so much change at once.
When the internet emerged, financial services didn’t jump straight to mobile apps and paperless statements. Many are still catching up, despite the technology being around for years.
AI will follow a similar path. Adoption in financial services will be evolutionary. Not because the technology is limited, but because organisations struggle to change. They struggled with digital transformation. Autonomous transformation is a step harder again.
Closing point
The firms getting value from AI share a few traits. They understand their processes. They take incremental steps. They learn from failure without blame and apply those lessons quickly. They also understand the real cost of adoption, not just the tooling.
They keep humans where it matters. Where judgement, context or client interaction creates differentiation.
Remove that, and you’re just another provider competing on price.
Chris Moore
Head of Solution Architecture
