Opinion · Orchestration

The Automation Trap: Why APIs and Workflows Won't Fix Fund Operations

If you ask ten COOs in fund services what their digital strategy is for the next twelve months, nine will give you the same answer: automation. Here's why most of them will be disappointed.

menu_book Start reading
Jonty Hurwitz
Written by
Jonty Hurwitz
Founder
Read time
8 min
Published
Recent

They want to scale AUM (Assets Under Management) without scaling headcount. They want to integrate their legacy ledgers with their CRM. They want to deploy AI agents to handle the avalanche of unstructured documents that define private markets.

But when these mandates are handed down to operations and engineering teams, projects stall. Budgets bloat. Shadow IT proliferates. Why? Because the financial services sector is fundamentally confusing three very different technologies: APIs, Workflows, and Orchestration.

Treating these three as interchangeable is the root cause of "automation debt." To build an operating model that can handle the complexity of modern fund services — where compliance, human judgment, and AI must intersect — we have to understand the difference.

01 · The API

The Fast (But Forgetful) Messenger

An API is a bridge. It allows System A — say, your KYC portal — to talk to System B, your core ledger.

The reality
APIs are brilliant for instantaneous data transfer, but they have zero memory and zero context. An API fires and forgets. If a call fails because a server times out, the data is dropped unless an engineer has custom-built a retry mechanism.
The fund ops problem
Fund operations rarely happen in milliseconds. They happen over weeks. You cannot run a 45-day capital call or an ongoing AML investigation purely on point-to-point APIs. It turns your architecture into brittle spaghetti that requires a developer every time a business rule changes.
02 · The Workflow

The Rigid Factory Line

Workflow automation — think traditional RPA or tools like Zapier — is a linear set of instructions: if this happens, do that.

The reality
Workflows are great for deterministic, low-stakes tasks. If an LP emails this inbox, save the attachment to SharePoint.
The fund ops problem
Fund administration is almost never linear. What if the LP sends the wrong document? What if a compliance check raises a red flag? Workflows break on the first exception, and they're notoriously bad at handling humans in the loop — the process times out waiting for approval and loses all context.
03 · Orchestration

The Stateful Conductor

While an API moves data, and a workflow moves tasks, an Orchestrator manages state, systems, and people over time.

Orchestration acts as the central nervous system. It knows exactly where a complex process is at any given moment. If a process requires pulling LP data via an API, waiting three days for an AI agent to parse a 100-page prospectus, pausing for a human Compliance Officer to execute a Maker-Checker approval, and finally pushing the verified data into an ERP — an Orchestrator manages that entire lifecycle end-to-end.

An API moves data. A workflow moves tasks. Orchestration moves state, systems, and people over time.

04 · The Next Matter approach

Why fund services needs orchestration

In private equity, venture capital, and asset management, the stakes are too high for basic workflows. Global players like Ocorian and Trade Republic don't just need systems to talk to each other; they need to govern highly regulated processes across thousands of employees and billions in AUM.

This is why we built Next Matter specifically as an orchestration layer for fund services. When you choose Next Matter over a generic workflow tool, you get an architecture built for the reality of financial operations:

groups
Human-in-the-loop by design
A GP signing off on an IC minute or a compliance officer reviewing an exception is a first-class feature, not a roadblock.
smart_toy
Agentic orchestration
Run custom AI agents inside a walled VPC alongside your human team. The orchestrator delegates, verifies, and hands off for approval.
verified_user
Auditability & governance
Every API call, every AI inference, every human click — logged. Maker-Checker is enforced by default.
05 · The bottom line

Three things to take away

The Spaghetti
If you run engineering
Point-to-point API scripts are brittle. Every business-rule change becomes a developer ticket. Orchestration centralises the logic so your team stops babysitting failed jobs.
The Time
If you run operations
Workflows are built for seconds. Real fund processes — onboarding an LP, closing a fund, running a capital call — happen over days and weeks. That's what orchestration is built for.
The Audit
If you own risk & compliance
Workflows don't remember. Orchestration does — every step, every decision, every approver. Hand regulators a link, not a war room.

You cannot build a scalable, regulator-ready fund operations team on point-to-point APIs and brittle workflow scripts. Scaling AUM without linearly scaling your headcount requires a platform that understands the nuance of long-running, complex, and highly secure operations.

Stop building factory lines for processes that require a symphony. It's time to orchestrate.

Keep reading

See what orchestration looks like in your stack

A working session with our team, on your real processes.