The middle office,
rebuilt for the AI era.
We're a team operating out of Dublin and Stockholm, built from the inside of global custodians, investment banks, and financial technology companies. We are building the AI-native operations layer that investment managers have always needed but never had.
We are looking for three design partners to build this with us.
The middle office runs on expertise that isn't captured,
data that isn't connected, and effort that doesn't scale.
Annual attrition of 25–35% in business process operations means the team running your middle office is perpetually rebuilding its knowledge base. This isn't a reflection on individual effort, it's a structural feature of the model. The drift is always in the same direction: slower resolution, rising management overhead, and a service that's harder to hold at consistent quality than it should be.
A shallow knowledge base, renewed on repeat.
High attrition means the team is always rebuilding. The knowledge that matters most isn't documented anywhere: why a counterparty behaves a certain way, how to read a pattern in aged breaks. It walks out with every departing analyst, taking resolution speed with it.
Break management is a full-time overhead.
Correctly allocating breaks to the right team, ensuring they've been actioned, and chasing resolution requires constant manager oversight: multiple people, every day. The coordination effort is significant, and it still doesn't always work.
Slow, inconsistent delivery at scale.
When expertise is shallow and processes are manual, quality varies day to day. Reconciliation reports arrive late. Scaling the headcount doesn't fix the underlying problem. It compounds the management burden without improving the outcome.
Data lives everywhere except where you need it.
Information is fragmented across emails, Teams messages, matching platforms, trade blotters, and broker and custodian statements. There is no single operational view of the business. Problems surface late, and decisions get made on incomplete information.
Reporting is manual, bespoke, and slow.
Because the data is fragmented, every meaningful report is a custom build. Bespoke client or management reporting requires manual assembly, cleaning, and formatting. Error-prone and impossible to scale.
An AI-native middle-office
operations layer.
The current model depends on people to hold knowledge, people to route and chase work, and people to stitch together data from disconnected systems. We replace those dependencies with an AI layer that retains institutional knowledge, connects your operational data across every source, manages break allocation automatically, and delivers consistent quality regardless of team turnover.
Reconciliation & Exception Management
Cash, position, and transaction recs across custodians and prime brokers, with AI-assisted break triage, root-cause classification, and automated allocation to the right team. The AI learns your counterparty patterns, custodian quirks, and break signatures over time. Knowledge that used to walk out the door now compounds in the system. Break aging drops. Manager overhead drops. Delivery arrives on time, every day.
Trade Settlement & Lifecycle
End-to-end trade lifecycle management from confirmation through to settlement. AI-powered matching, status tracking, and exception handling across custodians, brokers, and clearinghouses. One view of your settlement pipeline. Not assembled from three platforms and an inbox.
Collateral Management
Real-time margin monitoring, CSA optimisation, and collateral tracking across prime brokers and custodians. AI-powered exposure analysis and automated margin call workflows. A live view of your collateral position. Not assembled at the end of the day.
Operational Visibility & Reporting
Because the AI layer connects your matching platforms, custodian statements, trade blotters, and broker communications, meaningful reporting becomes a byproduct of operations, not a separate manual exercise. Reports that used to take hours to assemble are available in real time. Bespoke client reporting without the bespoke effort.
We are pre-launch and building. Design partners shape the product.
We're not looking for customers yet.
We're looking for partners.
A design partner works with us directly during build. You get: a product shaped around your actual workflows, priority access at launch, and commercial terms built for early movers.
We are onboarding three design partners. If you run a middle-office team and are frustrated by break volumes, manual settlement tracking, or collateral headcount — we want to talk.
No deck. No sales pitch. We'll ask direct questions about your operations and be honest about what we're building.
Built from the inside.
We are a small team based in Dublin and Stockholm. Between us we have operated middle-office functions at some of the world's largest custodians and investment banks, led AI engineering organisations at global scale, and built financial technology companies from zero in Scandinavia.
We are in stealth. We're not ready to introduce ourselves publicly — but we are ready to have a direct conversation with the right people.
Three forces converging — right now.
T+1 Settlement is live.
Settlement cycles are now live in the US, and coming shortly in the UK, and EU markets. The window to catch, investigate, and resolve errors before they become failures has halved. Manual workflows built for T+3 are visibly failing.
AI capability has reached production quality.
Large language models and structured data agents can now handle exception classification, break triage, and workflow routing at the quality level investment operations demands. This wasn't true three years ago.
Fee compression has made the status quo unsustainable.
Average management fees are down 40% over a decade. Middle-office cost structures built in the 2000s are no longer viable. Automation is the only lever that doesn't sacrifice quality.
Let's talk about your operations.
A 30-minute conversation. No deck from our side. We'll ask direct questions about your operating model and be direct about what we're building and where we are.