Over the last few months I have significantly changed my mind about how disruptive AI will be within financial services. Many long conversations with Thomas Byrne, our Head of Product, have shaped this thinking. What shifted my thinking is understanding what it really means to deploy an AI Native Operating System.
I used to believe that long-term value could be delivered by deploying agentic capabilities alongside your existing tech stack and processes. I no longer believe that. Deployed that way, agents deliver short-term value only. You are speeding up old ways of working, not changing the work itself.
But here is the nuance: in many cases it is exactly the right way to start.
Start alongside the stack
At Covecta we are delivering significant value for our clients today, and we are doing it without touching the core. Here is our current approach.
Our agents perform the manual work that sits outside of systems. Every bank, lender, credit union, building society runs on work that never shows up in a system of record: reading documents, assembling evidence, chasing information across teams, rekeying data from one screen into another, writing memos, reconciling financial performance. That is where our agents go to work first.
Our clients are typically live within 12 weeks. No disruption, no migration, no core replacement programme. Low risk, with high value. It is the fastest credible route to real outcomes, and it builds the confidence and capability an institution needs for what comes next.
That is the point most AI strategies miss. You do not have to choose between quick wins and structural change. You start at the edge, prove value in production, and innovate from the outside in.
The shift we see coming
Within 6 to 12 months we expect the centre of gravity to move. The institutions that started at the edge will begin to redesign what sits behind it. Previously to innovate you needed to have every detail ironed out before you committed to a technology migration project. Times have seriously changed, and technology decisions will too.
Your customer data gets unlocked. It will be housed in a centralised datalake: Databricks, Snowflake, Microsoft Fabric etc. Data will now be accessible to your mission-critical agents and no longer trapped inside legacy SaaS systems of record.
Processes get redefined for an agentic era. Rekeying of information across systems will be gone. Manual handoff points across teams will be eliminated. The process is designed around what agents can do, with humans on the decisions and judgement that matter.
Legacy systems get decommissioned. As they come up for renewal you will switch them off rather than roll them over, unlocking budget and reducing the complexity of your tech estate. You will give your incumbent providers the chance to prove they can innovate before the next renewal, if not then time is up unfortunately.
Work consolidates onto your AI Native Operating System. Agents and humans working in harmony on one platform. Not a copilot bolted onto old software, but the place where the work actually happens.
The place of work
This is where Covecta is heading. We will be the place of work for all your Lending, Account & Deposit, and KYC/AML & Compliance needs. Purpose-built for financial services, governed and auditable by design, with a flexible deployment model. Technology that you can rely on for the next century.
The sequencing matters. Institutions that wait for a perfect end-state architecture will still be writing strategy papers while their competitors are live. Institutions that stop at the quick wins will find the value plateaus. The winners will do both: start alongside the stack, then use that momentum to rebuild the work itself.
Why this matters
This approach is how you stay ahead of your competition. It is how you create a place of work your teams actually enjoy, because the rekeying, the chasing and the swivel-chair work are gone. And it is how you deliver more value for your customers, because your people spend their time on judgement and relationships rather than administration.
Deploying agents alongside your stack is the start. The AI Native Operating System is the destination. The institutions that understand the difference will define the next decade of financial services.
