B-Hive FinTech Hackathon

Insurance fraud is a phenomenon occurring within every stage of the insurance lifecycle: At policy subscription, giving fake identity data and fraudulent objects are among the many strategies that we see applied today. But also the stages of claims declaration (overclaiming, fake identities) and payout processing (identity theft, interception of payments) are not without threat.

Notwithstanding today’s technological advances, capabilities to detect such fraud attempts remain vastly unexploited. The ways in which we gather, use and share relevant data within and across industries remains a key obstacle for fraud detection strategies to reach their full potential.

So how might we effectively detect fraudulent claims using innovative approaches such as AI and natural language processing?

We are not looking for a “one size fits all” solution. Every form of fraud detection is different and may require a specific solution. Combine all your technical expertise and ingenuity to uncover a variety of solutions that could help in this fraud detection process.

Can you come up with a solution to:

- gain additional intelligence within large datasets (structured and unstructured data) to automatically flag suspicious claim declarations (pattern detection)?

- detect fraudulent claim submissions using pattern detection across insurance policy datasets?

- enrich the fraud detection system with external datasets? They can vary from open source datasets to other private datasets.

- use unstructured datasets (such as claims descriptions, voice to text from calls, …) to detect possibly fraudulent claims?

Your solution must be applicable in at least one of these insurance domains:

- Car Insurance

- Home Insurance (theft, fire, natural disasters…)

- Labour/Liability Insurance/Guaranteed Income

- Life/Health Insurance

Procurement processes are often (rightfully) long and tedious for both parties involved. The corporates have heavy, often highly manual processes to screen the companies, while the tech companies have to go through this process for each customer again and again. In the current environment, speed of innovation is key, and API based systems facilitate a faster “plug and play” integration. Hence, the procurement processes need to be reviewed and adapted for this new challenge. As such, we need to find enablers to help speed up the entire process, while ensuring the right level of security/risk at each moment. We could, for instance, define a step by step approach, gradually increasing the level of detail of the assessment, linked to the phase of engagement (eg: Basic screening for a POC, deeper screening for pilot and full screening for full implementation tracks, etc).

In this challenge, we want to chop up the process in manageable pieces, and try to find a technology driven solution to assist or speed up the process for each part of the procurement journey:

Can you come up with a solution for a generic KYV platform for collecting, sharing and updating vendor-related procurement data on a cross-industry basis (so as to speed up the procurement process and avoid duplicate work)? The idea is not simply to outsource this to a centralised provider but to create an API based system where individual companies are able to create their own tailored systems customised to their own risk appetite.

- Submission of company information: how to use external datasets or analytics to get a validated 360 view of the company (shareholder structure, funding, team, existing customers, …)

- Financial analysis (how healthy is the company?): can we use A.I. or external datasets to speed up the financial assessment of companies?

- Risk & security assessments: tooling to assess basic and advanced security:

• penetration testing
• API security
• ...

- (Regulatory) Compliance checks:

• GDPR compliance
• Any other regulatory compliance (eg PSD2, IRFS9, …)
• …

Belgium is well-equipped when it comes to e-payments. There is great infrastructure in place and most Belgians do use electronic channels. However, given recent technological advances, there is tremendous opportunity to increase adoption of mobile payment solutions and generally provide a better user experience. We have to keep in mind, however, that such improved services must be accessible to all, including youth, students, elderly people, families, physically impaired people, and the currently unbanked such as refugees and other disadvantaged groups

So how can we make sure that we are fully inclusive with our mobile payment solutions?

KEY WORDS: Electronic Payments, Reaching the unbanked, Financial Inclusion, Loyalty, Local Merchant solutions

SMEs and independent professionals face a tremendous administrative burden with regard to the management of their finances: think invoicing, payment, expense management, accounting, financing, and managing cash flow and liquidity. While there are many solutions available in the market that solve individual aspects of this wider challenge (such as online accounting software), the fragmented nature of the market makes it difficult for banks to play a more proactive role supporting their customers.

Ideally, banking services should be integrated in end-to-end financial management platforms, so that more tasks can be automated, more data can be leveraged for better insight and prediction, and liquidity is always optimised. In this challenge we are looking for solutions that could help banks serve their SME and professional customers better, especially with regard to:

• sending/receiving the invoices digitally
• facilitating the decision to pay the invoice
• initiating payments
• tracking open invoices and payment status
• optimizing expense management of employees
• keeping a real-time overview of the cash flow and liquidity
• getting financing
• optimising working capital needs
• decreasing Days Sales Outstanding
• predicting and solving liquidity issues
• connecting to tax and accounting platforms

The goal of this challenge is to detect fraud in a transactional environment by combining and analysing multiple data types and sources. The Hackathon partners will provide you with anonymous transactions data (both fraudulent and clean transactions) and detailed descriptions of other transactions data, but your task is to propose a viable approach and method for combining and analysing data from multiple financial institutions and external, publicly available data.

You will work with the partner’s experts to address IT issues, data issues (including proposals on external data sources), security/privacy issues, analytical issues, and business model issues. To succeed, your proposed initiative will have successfully withstood technical and business scrutiny by the partners during the hackathon, meaning it could be viable and is thus worthwhile to explore further in a proof of concept.