Peter Barcak and the risk of saying no

Peter Barcak has a joke about his former profession. Risk managers, he says, never make mistakes. The correction arrives immediately: of course they do. It is an economical introduction to someone who spent more than two decades making decisions about other people’s borrowing before putting his own career on a different sort of balance sheet. In 2016, he moved to Singapore and launched Credolab. The familiar business of evaluating risk had acquired an unfamiliar applicant: himself.

He brought a lender’s frustration with him. As a chief risk officer, Barcak had found that the information available to make a credit decision could be painfully incomplete. There were people a bank could not understand well enough to serve. A rejected application closed the matter for the institution. For him, the unanswered question remained: what useful evidence had never made it onto the desk?

The application form was too small

His working education came inside banks. Citibank supplied formative years. At Intesa Sanpaolo in Europe, he helped build retail risk management from scratch, establishing policies for decisions on products such as mortgages and credit cards. The work demanded rules that other people could apply repeatedly, with real money riding on the result. A clever argument was of limited use unless it survived the next application.

Then came Platinum Bank in Ukraine, where he served as chief risk officer and deputy CEO. He has described it as his first startup experience. There, the distance between a customer’s circumstances and the bank’s information became a business problem he wanted to solve. Around 2013, he began experimenting with alternative inputs, including telecommunications and social media data. The eventual company grew out of that search, rather than a sudden conversion to fashionable technology.

A conventional application can offer only a modest collection of fields. Barcak wanted to know whether the device already in a customer’s hand could add something useful. That question required its own discipline. A large quantity of information is easy to admire. Finding the small amount that improves a lending decision is a more demanding occupation.

“The hardest thing for a risk manager is probably to take risks yourself.”

Peter Barcak, December 2022

Singapore, with the risk in his own name

Barcak has called the move to Singapore and the launch of Credolab the turning point in his career. It gave him a new market and a chance to build around the problem he understood. His background offered a practical advantage: he could talk to potential customers as someone who had sat on their side of the table, responsible for a loan portfolio and its consequences.

By 2018, he and Michele Tucci were presenting CredoScore at FinovateAsia. The early public demonstration described a product that collected smartphone data with permission, processed it into a score and worked alongside a lender’s existing risk processes. The audience was being asked to consider a new input to a familiar decision. The bank would still have to decide what to do with it.

The presentation also makes the partnership visible. Tucci, then chief product officer, stood alongside Barcak in explaining the product. Building a company around lending data required more than a founder’s conviction. Someone had to turn the argument into a usable service, demonstrate it and explain how it would fit a customer’s working day. There is a pleasing lack of romance in that requirement. The software has to arrive before the speech can matter.

A few dozen signals, chosen carefully

Barcak’s explanation of the technology begins with metadata: information about activity rather than the contents of a personal message. He has used examples such as application categories, calendar activity and device usage. The company looks for combinations of behavioural features that correlate with outcomes and remain useful over time. No single habit provides a miniature biography of a borrower.

He describes behavioural assessment as a complement to conventional credit information. A lender may know something about someone’s ability to repay and still want additional evidence about likely behaviour. His company’s job is to supply another predictive input. In that account, an algorithm adds to the decision-making process while the lender retains responsibility for the decision.

The distinction also helps explain his insistence on permission and the handling of data. Credolab says it uses consented metadata and avoids collecting personally identifiable information. Those promises belong beside any explanation of the product, because the proposed benefit depends on people allowing software to examine signals from their devices.

The discipline of selection was present in the early product. In a 2020 account of the technology, Barcak explained that a scorecard usually contained around 20 to 25 features with high information value and predictive power. The huge pool of possible inputs was a starting point. What reached the scorecard had to earn its place. A risk officer could recognise that habit immediately.

A company has to pass its own test

In November 2019, Credolab won the ASEAN FinTech award in the Open category, presented by the Monetary Authority of Singapore and the Association of Banks in Singapore. The prize came with S$150,000. For a company selling a different way to evaluate borrowers, recognition from the financial establishment was a useful development. The proposal had moved beyond a product demonstration and into an industry’s public conversation.

The next year brought a larger financial milestone. Credolab announced a US$7 million Series A led by GBG, with Access VC and existing investor Walden International participating. The stated plans included strengthening its presence in Asia and expanding into Latin America and Africa. The round supplied resources for an international business, along with the obligation to make that expansion work.

Expansion did not always require the travel schedule he might have expected. Barcak has recalled receiving enquiries from across the world and signing business remotely from Singapore. That experience helped shape a model of small teams and regional centres. A software company could reach an overseas customer without first reproducing itself, office by office, across a map.

The company’s later financial choices added another test. In 2025, Tucci described a shift towards profitability, lower cash burn and growth funded by the business. He said Latin America had become a major contributor to revenue. For Barcak’s company, financial inclusion remained the customer proposition, while the internal work involved the ordinary demands of running a solvent enterprise. Both sets of numbers required attention.

The company he keeps

There is a short sentence in Barcak’s advice about succeeding in finance that deserves to survive the surrounding jargon: “Be kind.” He argues that understanding people and treating them well matters in business. From someone whose profession involves pricing uncertainty, it is refreshingly direct advice. No diagram is necessary.

A former colleague remembers a similarly direct question. In a public recommendation, Vadym Kryvovyaz recalls Barcak asking, “Are you a happy person?” during their first job interview. Kryvovyaz went on to describe seven years working with him and a progression from risk analyst to chief risk officer at Platinum Bank. The recollection gives a more particular glimpse of management than a list of leadership adjectives could provide.

Outside work, Barcak has spoken about the pleasure of discussing business and the wider world with his three children. He enjoys their curiosity as they grow old enough to engage with subjects he has spent years thinking about. It is an appealing reversal of the executive interview: the experienced founder gets an audience whose interest cannot be taken for granted.

His entertainment choices include Mr. Robot and the 1984 film Amadeus. His preferred ice cream is chocolate. None of these details needs to become a theory of entrepreneurship. Chocolate has already endured enough marketing. Together, they give the person some room beside the job title: a parent who likes a conversation, a film and a dependable scoop.

Getting inside the next decision

The problem Barcak began with keeps producing adjacent questions. In November 2025, Credolab announced an Income Prediction Model intended to estimate income from consented smartphone metadata when conventional documentation is unavailable. It was initially available for Android devices. The company describes models customised to a lender’s data and local population. The underlying ambition is recognisable: help assess an applicant whose paperwork leaves too much unexplained.

Another route opened in August 2026, when Credolab announced its partnership with FICO Marketplace. FICO Platform clients could access its behavioural risk scoring and related intelligence within existing workflows. For Barcak, who has recently described distribution as the next phase of the business, the practical significance is clear. Useful information has to reach the place where someone makes a decision.

That concern runs through his career. He first built the rules used inside banks. He then questioned whether the evidence feeding those rules was sufficient. Now he is working to get another kind of evidence into lenders’ hands. The distance between those jobs is considerable, but the question connecting them is plain: before declining a person, what else should a bank know?