Macgregor Duncan and the Bank You Never See

Before Macgregor Duncan asked banks to rethink their machinery, he helped convene a gathering with a pleasingly unbankable name: the Australian Half-Baked Ideas Seminar. At Harvard in 2002 and 2003, he and fellow Australians met to discuss the country’s future. The name left room for an idea to arrive imperfectly dressed. Eventually, the discussions became a book.

Imagining Australia: Ideas for Our Future appeared in 2004, written with Andrew Leigh, David Madden and Peter Tynan. Its subjects ranged from national identity to democratic institutions and economic growth. The book looked ahead to 2020. Duncan has since moved from proposing a future on paper to building a business with rather more immediate deadlines.

Today, he is co-founder and co-CEO of Constantinople, alongside Dianne Challenor, known as Di. Their Sydney company supplies software and operational services to banks. Much of its work sits behind the familiar business of opening an account, making a payment or applying for a loan. The customer may never encounter the company’s name. Duncan has chosen a consequential place to be inconspicuous.

A book written between places

The early writing project was geographically untidy. Duncan described collaborators spread across cities and countries, working through email, instant messages and telephone conferences, with occasional meetings in person. A book about a shared national future had to survive the ordinary difficulty of getting four people into the same conversation.

His interest in public affairs went back to high school. He graduated from Adelaide with first class honours in politics and law, then worked for Justice Michael Kirby at the High Court of Australia. Princeton followed, with study in public policy and international relations, and then a master’s in law at Harvard. By the time the book appeared, he was working at a New York law firm.

The book’s table of contents gives some sense of the territory its authors wanted to cover: Australian identity, democratic reform, nation building, prosperity and the country’s place in the world. It was a large agenda for four collaborators with day jobs. They put their proposals in print anyway.

There is a useful connection between those interests and Duncan’s current work, although the projects are separated by very different careers and circumstances. Both concern institutions that people depend on. Both ask how an inherited arrangement might work better. A seminar allows a proposal to remain unresolved until next fortnight. A banking platform has to process the next transaction.

The long route to a shared problem

Duncan’s professional itinerary resists the tidy founder biography. He worked as a corporate lawyer at Debevoise & Plimpton in New York and as an investment banker at Goldman Sachs. He led global corporate development at Better Place, the electric-car infrastructure company. His experience also includes a managing director role in New York private equity.

These were different kinds of organisations, with different things to sell and different ways of getting paid. The work placed him around transactions, capital and the practical problems of growing a business. His older author biography also records writing on education, democracy and innovation. The banking founder had other subjects on his desk.

He joined Westpac in September 2016. His corporate and business development responsibilities included acquisitions, strategic investments and partnerships. He sponsored the bank’s Co-Lab and FuelD data accelerator and sat on the investment committee of its venture fund, Reinventure. Later, as Chief Development Officer, he helped launch Westpac’s digital bank.

Westpac was also where he and Challenor worked together before founding Constantinople in 2022. Their subsequent venture addressed a problem they had encountered from inside banking: the difficulty of coordinating the technology and operational work beneath a financial service. They had experience of being the institution that needed a solution. Now they would have to persuade other institutions to buy theirs.

All the things behind the account

When Duncan explained Constantinople in 2023, he made its relationship to banking clear: “We don’t run a balance sheet.” The company sells services to banks. A client’s licence, balance sheet and customer relationships remain central to what that bank does.

The proposed division of labour is broad. Constantinople’s platform covers banking products and customer interfaces as well as infrastructure, operations, servicing and compliance. It can handle work such as customer onboarding, fraud monitoring and payment operations. Rather than commissioning each component separately, a bank can buy a managed service that connects them.

That breadth helps explain the challenge. An attractive account screen is easy to understand. The operational work behind it is harder to display, and a prospective customer must be comfortable with the company doing it. During the public launch in May 2023, Great Southern Bank was the first client, preparing a staged pilot for its business banking service.

The early backing came from Square Peg, Airtree and Great Southern Bank, with an announced A$32 million seed round. In April 2024, a further A$50 million Series A was announced, led by Prosus Ventures. Great Southern Bank’s business bank was live. The money funded more work; the live service gave the proposition something a funding announcement alone could never supply: a customer using it.

The presentation had competition

Trust became particularly tangible in the relationship with MoveBank. Its CEO, Mike Currie, recalled visiting Constantinople with his chairman. There was a substantial presentation. There were introductions to employees. Then a staff member showed a video of colleagues playing basketball before work.

Currie remembered the video as evidence of the team’s culture. It is a wonderfully inconvenient detail for anyone hoping the slide deck did all the selling. He was assessing people who would become part of the bank’s daily operating life. An informal glimpse of those people stayed with him.

MoveBank subsequently became Constantinople’s first full retail-bank migration. Duncan announced the move in August 2026. In his account, the scope included customers, products, channels, operations, compliance, security and reporting. He also reported a reduction of more than 30% in MoveBank’s cost-to-income ratio. That figure is the company’s account of the outcome.

For Duncan, the wider argument is about access to scale. A smaller bank can remain a separate institution while sharing a platform with others. He calls that “synthetic scale”. It gives his company a proposition beyond a technology replacement: banks could consider a shared operating platform alongside the familiar option of merging with another bank.

“There is no margin for error.”

Macgregor Duncan, on handling banks’ financial data and operations, 2023

The quiet argument inside the AI argument

By January 2026, Duncan was writing about what he thought the next phase of banking technology would require. One essay argued that shared infrastructure platforms could change the economics of incumbent banks, including smaller ones. His attention was on operating costs and the organisation of work underneath the customer experience.

A companion essay examined why AI agents struggle in banks. Duncan’s diagnosis began with fragmented data: separate systems can carry inconsistent definitions and incomplete information. He also argued that recording a decision is insufficient if the reasoning behind it disappears into conversations and inboxes. An agent needs context for the exceptions that make institutional work complicated.

“You can’t turn dust into gold,” he wrote. His proposed foundations include a consistent data model, records of decisions and controls over execution. Approvals, accountability and the ability to correct mistakes belong inside the design. These are his arguments about what banking AI needs, rather than a promise that every part of that future has arrived.

The concern with control was present earlier. In 2023, Constantinople was using GitLab to build security checks into software development, with automated processes and frequent releases. Duncan stressed the responsibility involved in handling sensitive financial information and essential banking operations. Speed has to coexist with the consequences of getting a transaction wrong.

A Roman name, an Australian ambition

In explaining the company’s name, Duncan reaches back to Constantine’s relocation of the Roman capital. The lesson he draws is that an institution can reach a point where repeated adjustments are insufficient and a new foundation becomes attractive. It is an expansive historical analogy for a company whose everyday work includes payments and regulatory reporting.

His geographical ambition is more contemporary. Duncan has spoken publicly about wanting to build a global technology company from Australia, creating employment and keeping the business headquartered and taxed there. That returns him to a subject visible in the book two decades earlier: what the country can make, and what conditions allow it to make more.

There is also a life beyond the platform. His Westpac biography lists hiking, camping and fishing among his interests. It is a welcome change of scenery in a career otherwise populated by institutions, investment committees and systems. Some activities permit a much thinner agenda.

The old seminar and the present company offer two different measures of an idea. One produced a book that readers could argue with. The other has to earn a bank’s confidence and keep earning it in operation. Duncan’s current version of imagining the future comes with customers, colleagues and a payment that needs to arrive.