The future of finance: AI that elevates people, not replaces them
By Kumar Ujjwal, Founder and CEO
The Real Question Leaders Are Facing
A few months ago, I was sitting in a conference room with the COO of a mid-market credit fund.
He had a stack of physical binders in front of him: covenant reports, capital call notices, side letters, emails printed out and tabbed. He tapped the top binder and said, half-joking, half-tired,
"This is what my team's talent gets spent on."
These are some of the most capable people in our industry. Many trained at global banks, survived multiple cycles, and have stories from '08 that younger staff listen to like folklore. Yet a shocking amount of their day is still spent hunting through folders, reconciling numbers between systems that don't talk, and triple-checking spreadsheets at midnight.
That's the part nobody puts in the marketing deck.
What "Experience" Has Quietly Become
When we say "this person is indispensable," we often mean:
- They know which version of the spreadsheet is the real one.
- They remember why a particular LP always needs a slightly different template.
- They've seen the reconciliation break in exactly this way before, and they know the fix.
That is real experience. It's valuable. But it's also fragile. It lives in a handful of people's heads and inboxes.
One fund administrator we spoke with had a senior operator who, frankly, held the whole shop together. When she went on a three-week leave, the team realized that critical parts of their institutional knowledge were only documented in her memory. Nothing "failed," but everything slowed down, and stress went through the roof.
The lesson wasn't "she's replaceable." The lesson was harsher: the system around her had failed her.
A Different Kind of Story
Now contrast that with a smaller, newer firm we work with. They don't have the biggest brand. They don't have a massive team. What they decided they would have is a different operating model.
Their junior analyst still builds capital call packages, but the first draft is generated by a workflow that reads underlying documents, pulls the right data, and assembles it in the correct format in hours instead of days.
Their senior controller still signs off on reconciliations, but instead of rebuilding checks from scratch each month, she reviews exceptions surfaced by a system that has already done the heavy lifting.
No one's judgment is being replaced. But their judgment is now applied where it matters:
- Understanding edge cases and portfolio nuance.
- Talking to LPs when something is off.
- Thinking about risk, not wrestling with CSV files.
- Building relationships that drive the business forward.
If you walk into their office during a busy close, you notice a different energy. Less panic. Fewer "which file is the latest version?" questions. More time spent in actual conversations about the portfolio. More time spent on the work that only humans can do.
The Real Question Leaders Are Facing
So when people ask me, "Will AI take jobs in investment banking and fund administration?" I think about those two rooms.
The existential risk is not that a model writes a memo or reconciles a dataset. It's that we keep asking humans to do work that systems are now objectively better at — and then wonder why we struggle with retention, morale, and margin pressure.
LPs are not getting more patient. They want transparency they can trust, faster than before. Regulators are not lowering the bar. Fee compression is not easing up.
In that environment, the real leadership question is:
Are we building an organization where our best people spend their time on the highest-value problems — or on compensating for the gaps in our systems?
That's not a "junior versus senior" question. It's an operating model question.
What We're Building at DwellFi
At DwellFi, we think a lot about this idea: a machine that builds machines — infrastructure that captures institutional knowledge and turns it into repeatable, auditable workflows.
In practice, that looks like:
- Turning institutional processes into AI-native workflows. That reconciliation process a senior operator perfected over a decade? It now runs the same way, every time, without human intervention on the routine parts.
- Encoding the nuance of LP reporting. Those requests that used to be "tribal knowledge" — the specific formats, the edge cases, the relationship quirks: become logic the system understands and executes consistently.
- Accelerating the ramp for new talent. A junior analyst can now stand up a deal room, populate it with clean data, and generate outputs in days — without needing three senior people to show them "how it's done here."
- Freeing senior operators from the grind. Your best people stop being the system's workaround and start being the system's brain.
When this works, something subtle but profound happens.
The junior hire ramping up doesn't have to be a hero. The senior leader doesn't have to be a single point of failure. The organization is no longer powered by individual endurance, but by a system that respects people's time and attention.
The Firms That Will Win
That's the future of finance I'm interested in. Not a race to replace people, but a deliberate effort to build systems that are worthy of the people we already have.
The firms that win the next decade won't be the ones with the most people. They'll be the ones where:
- Technology and human judgment are not in tension: they're designed to amplify each other.
- Institutional knowledge is captured and scaled, not hoarded by a few.
- People are freed to do the work only they can do: thinking, judging, building relationships, managing risk.
- Speed and accuracy improve together, not at the expense of one another.
If you're a COO, CFO, or Managing Director wrestling with these questions, I'd be glad to compare notes. We're still learning every day, but we're committed to one simple idea:
The future belongs to the firms that build systems worthy of their people.