DigFin has ranked the presentations of five fintechs that have made the finals of an annual incubation program run by Accenture.
Eight startups reached the finals, but DigFin has already covered three: BlueOnion, Evident and Planto. Here are links to each company’s founder in discussion on our video podcast, DigFin VOX:
Okay, and then there were five. Based solely on their brief pitches, here is how DigFin perceives their business model’s potential, with some questions that investors or potential partners should bear in mind.
5. Oxford Risk (UK)
This Oxford University spinout has developed software to help financial institutions understand how people make financial decisions. Better decisions can ensure retail investors grow assets by being exposed to the right products, instead of remaining too heavily exposed to cash. The company calls such insight “behavioral alpha”.
The startup began in the wake of the 2008 financial crisis with the insight that wealth managers (banks, financial advisors) struggle to understand clients, while retail investors are intimidated by jargon. One’s ethnicity, gender or other traits matter less than their understanding and appetite for risk, and how comfortable they are with such questions.
The company says it is now integrated within the tech stacks of multiple financial institutions across 15 markets.
DigFin’s interpretation of this business is that it is grounded in psychometrics, a term not raised in the pitch. The company’s methodology to gain insight is questionnaires. Such tools have been around for a long time, and are used by banks to help assess a borrower’s creditworthiness. They can be useful add-ons but don’t by themselves provide a complete client assessment. This is the first time DigFin has seen these tools applied to retail investing, but the principles are probably the same, as are the limitations – especially 25-question surveys.
4. Bilby (Hong Kong)
Bilby’s bots constantly scrape the internet for government announcements and related news. Then it applies predictive models, so its customers get an early heads up on what governments might do. Its pitch cited a bank’s prescient warning to clients about China’s sudden decision to crack down on its edutech sector. Such information could be parsed systematically.
Bilby’s founding team are capital-markets vets so they understand the use of plugging its data insights into the desks of quantitative investors and traders. The key to making it useful is to source lots of structured data, but that alone doesn’t add sufficient value: it’s the API into trading desks. They can receive signals around upcoming or real-time policy and regulatory changes, and then build trading models on top.
J.P. Morgan is trialing it. The team is based in Hong Kong but the aim is go global.
DigFin’s take from the pitch is that this is indeed useful, but how wide is the company’s competitive moat? Hedge funds have been buying or building sentiment-analysis software for years. Bilby may have found a new niche, but could a more established company simply add their idea to an existing suite? The maxim that ‘the proof is in the pudding’ applies more than ever to this service, because Bilby’s success will depend on traders attributing alpha to its insight.
3. Libertify (France)
This company transforms documents and data into virtual storytelling: think deepfakes, only legal and legit. It uses AI to provide financial institutions with a video experience: ‘video is eating the world’ is the company’s unofficial tagline. More clients and counterparties prefer video to reading, so market participants need the tools to communicate that way, while using AI-generated avatars to deliver the content.
Examples include relationship managers running clients through the week’s market performance and what’s coming up; or quick marketing turnarounds. The avatars are reliable (that is, they are compliant) and they are now of high quality. They are meant to be engaging: the AI is meant to scan a PDF and create a video that delivers the key information fast and in a friendly manner. There’s also a video-chatbot capability to answer questions, a useful means of extracting more data about client sentiment and concerns.
This is more active than simply emailing them information and FAQs, and if the tech can generate a two-way dialogue, there’s a better chance of nudging the client in a certain direction.
Libertify is now rolling out trials with Société Général for its listed products in Singapore.
DigFin acknowledges this sort of thing is going to become commonplace, but that it’s also harder than it looks, and so Libertify can gain competitive traction by pushing the envelope. Our question concerns how people respond to AI-only interactions: will they really engage, or will they tune out? These LLMs can get boring pretty fast, and there’s a risk that everyone ends up drowning in AI-generated sludge. The answer is to use this, like any other tool, judiciously, but startups need to scale, which implies a contradiction they and their clients will have to navigate.
2. zkMe (Hong Kong)
A new day, a new acronym: DAD, for decentralized autonomous data. Everyone in financial services wants structured (machine-readable) data, but this data is vulnerable to data-sovereignty laws and other frictions to its sharing and use. This is especially problematic in the emerging world of decentralized finance, which is meant to be global and accessible to all.
The implication of self-sovereign data is that identity becomes interoperable, and companies or individuals don’t have to worry about being liable for third-party data. It removes the rationale for data brokers.
Enabling digital identities that can travel seamlessly across markets and protocols will unlock plenty of commercial ideas for blockchain-based activities. For example, if people or companies have their own trusted ID credentials, they can access structured data from government apps or private wallets or a commercial bank – without needing to trust an AI agent or a broker, and without pooling everyone’s vital information in a single data base that could be hacked.
ZkMe wants to be the universal layer for identity, like a virtual biometric card using mathematical proofs that your ID has been verified without having to show the actual data. This can eliminate costs for due diligence, data liabilities, and still meet anti-money laundering and national data-protection laws. And data only gets shared when it must be, rather than sitting on third-party servers.
Although zkMe’s solution is now live in crypto markets, the startup is talking with commercial banks and telcos.
DigFin’s take: well, data sovereignty has been a dream, and a graveyard, for many a startup. It’s also a busy space, with lots of other teams gunning to produce a viable solution. DigFin doesn’t have the knowledge to compare one team’s tech to another’s. But the technology has advanced, and so has the need. Web3 is all fine and groovy, but the real use case is that we desperately require tools to restore trust in this new age of scary AI deepfakes and fraud. Solutions such as zkMe’s could be part of that package. If zkMe succeeds with bank and telco PoCs, either integrating them into blockchain commerce or using zero-knowledge proofs for existing business, there’s reason to be optimistic about this startup.
1. Otonomi (USA)
Usage-based data proved itself first in the retail world: think ZhongAn digitizing flight-delay insurance in China.
Otonomi is bringing the concept to the bigger, more complex world of marine and air cargo insurance. Every year some $24 trillion worth of goods moves by sea and by air, but cargo is often delayed – and delays are getting longer, thanks to factors ranging from droughts to congested ports and canals, to Houthi rockets. These delays strand 460 million containers at sea every year for a week or more.
While companies can insure products on the move against damages or losses, they can’t get delay insurance. Shippers are out of pocket. Otonomi reckons this represents a $50 billion protection gap.
It is now introducing delay insurance based on a single metric: transit time. If this exceeds six days by sea or three hours by air, the company will pay out, no questions asked, no claim filed. It does so based on sensors tracking cargo that integrate data with logistics companies via APIs, to enable data-driven predictions and contract pricing; plus blockchain to process claims. Otonomi practices as a licensed insurance agent.
The startup is now working with global insurance brokers such as Marsh, as well as cargo specialists, logistics companies, and freight forwarders. The bigger implications beyond insuring cargo delays is to help improve carriers’ underwriting and risk management, mitigate losses due to climate risk, and use parametric measurements to reduce fraud.
It is keen to expand to Asia and has set up a beachhead in Hong Kong.
DigFin’s question is the obvious one: if this is a good commercial business, why hasn’t anyone done it before? Surely a Lloyds name could have done this. Our second question is: how many players need to integrate with Otonomi’s APIs for this service to work? Looking at efforts to digitalize trade financing using blockchain suggests business models that rely on a network effect are very hard to sustain. So those are our questions from this pitch, but DigFin found the problem to be real and sizeable, the TAM kinda-sorta credible, and the tech aspects have been proven in other contexts. Usage-based data streams offer many improvements to current business models, provided the data is voluminous and the market size can scale – which is the case with Otonomi’s pitch. Moreover, if the startup can gain traction in this market, it will be positioned to tackle bigger parts of the P&C market.
