Aigang wants to use Ethereum smart contract technology to provide fully automated insurance products. The Aigang team, lead by Agustas Staras, plans on building a prediction market to crowdsource actuary and data modeling work. Aigang will use the insights from the prediction market to develop new insurance products, specifically targeting the market for consumer internet connected devices.
Aigang’s first product, currently available on Android and iOS and running on the Ethereum testnet, is an insurance plan for mobile phone batteries. For a monthly premium paid in ETH, users can insure their phone’s battery against normal wear and tear. The app hooks into the device’s API and monitors the condition and performance of the battery. When the battery’s performance dips below a certain benchmark, the insurance policy automatically pays out ETH that the user can put towards the purchase of a new battery.
Insurance policy terms are presented to the user in the form of a Solidity smart contract. Aigang has also left some questions about fraud prevention and dispute resolution unanswered. What is the potential that smart contract insurance policies could be manipulated or gamed? How will customer disputes be resolved? Given their pre-product stage of development, these are questions which the Aigang network is choosing to address at a later stage. In an interview, Agustas stressed that the team’s focus right now is on the development of their core technology.
What is the token being sold?
The Aigang Token is the AIX, whose sole purpose is for staking and rewards in the prediction market. It will not be a form of payment for Aigang insurance products. The primary currency of exchange on the Aigang platform will be ETH.
Prediction markets are still largely experimental economies. At the core of a functioning prediction market is the requirement that users back their predictions with a store of value, a bet if you will. This ensures that market participants have ‘skin in the game’ and incentivises them to answer truthfully.
Gnosis chose to use ETH as the staking currency in their prediction market. ETH is a relatively well known and trusted store of value within the crypto community. There is little doubt that users staking ETH have something at stake. This is less certain in the case of a new cryptocurrency. Since the Aigang token is not tied to the sale of Aigang’s insurance products or the performance of the platform, its initial value will come from the ability of Aigang to fund rewards for crowdsourcing competitions. Over time, if the Aigang prediction market proves to be a fruitful environment to develop and test actuarial models and insurance products, then the market for Aigang tokens may become more self supporting.
The details about staking requirements or reward schedules for the prediction market have not been defined in the whitepaper.
The project status
According to founder Augustas Staras, the Aigang app has been downloaded on over 8,000 devices since its launch in July, 2017. At the time of writing, over 2,000 users had tested out the phone battery insurance policy. This means that the Aigang network is collecting battery performance data from over 2,000 devices. This data will be used to build actuarial models and insurance products that will be evaluated in the Aigang prediction market. The models with the most support will be developed into insurance products and their creators and backers will be rewarded with Aigang tokens.
Aigang plans to use funds from the token sale to launch the phone battery insurance policy on the Ethereum live net and build a prediction market. The Aigang team opted to develop their own prediction market rather than build on top of existing protocols such as Gnosis. In an interview with the Smith + Crown team, Agustas cited their desire to customize their prediction market for their specific needs as the reason for not building on an existing prediction market infrastructure. At the time of writing, Aigang has not released any public code for their prediction market.
Their near term roadmap also includes the development of new policies for drones and Tesla vehicles. These are consumer devices, which like phone batteries, have internal APIs that can function as oracles to deliver trusted data about condition and performance to the Ethereum blockchain.
Who is the team behind the project?
CEO Augustas Staras a decade of experience managing and growing online businesses. He was also involved in the Mysterium Network crowdsale, which raised more than 14 million USD in it’s 2017 token sale.
The Aigang development team is lead by Darius Devans, a full stack developer with experience in smart contracts. He a former senior developer at Adform, a reporting platform for media agencies. The development team is advised by Bok Khoo, aka bokkypoobah, a high profile ethereum developer who has been involved in dozens of crowdsale audits. Bok Khoo also developed the code for MobileGo’s token, the ENS auction token, and OpenANX’s token.
|Incorporation status||Superposition Technologies Pte. Ltd., Singapore|
|Team openness||Fully transparent|
|Blockchain Developer||Darius Devenas|
|Technical White Paper||Available|
|Available Project Code||Insurance contracts and token sale contracts|
|Prototype||Available on iPhone and Android|
|Role of token||Participation in in-app prediction market|
|Distributed in ICO||51% of total supply|
|Emission rate||No new coins created|
|Consensus method||Proof of Work|
|Sale period||November 15 to December 15, 2017|
|First price||1 ETH: 2300 AIX|
|Investment Round||First public offering|
|Token distribution date||Unknown|
|Min investment goal||None|
|Max investment cap||45000 ETH|
|How are funds held||Smart contract|
|Minimal Viable Product||Q3 2017|
Blockchain and the Insurance Industry
The promise of blockchain-based insurance has captivated many entrepreneurs and insurance providers. Aigang will join a growing group of major players in the industry.
Blockchain technology has several applications in providing insurance. While they work well synergistically, they are not necessarily interdependent.
- Immutable Auditable Data
- Automated Claims Processing
- Distributed Risk Assessment
Immutable Auditable Data
The first application simply involves more efficient data storage and management. Distributed ledgers could be leveraged to store claim data and generate an immutable, auditable trail of activity throughout the lifecycle of a policy or a policyholder. Such a ledger could be helpful in identifying fraudulent or duplicate claims, fraudulent or duplicate accounts, lapsed payment or coverage, etc. The need to synchronize databases across the networks of insurance and reinsurance parties that represent modern insurance markets is a cost driver in the industry.
This application alone is promising but not necessarily disruptive, because much of the benefits could be reached through better data management practices in general. In addition, the full benefits of distributed data management for policies and policyholders rely on yet-to-be-realized advances in blockchain-secured identity and privacy. For example, healthcare insurance might be hindered by the sensitivity of patient data, which couldn’t be published by oracles onto an open publicly viewable blockchain.
Automated Claims Processing
The second application involves smart-contract based claims processing and has more exciting implications. Claims processing suffers from several inefficiencies, including manual review of claims and their supporting evidence and slow payment channels. Whenever an event that would trigger a claim occurs, the policyholder must prove that it occurred and the claims processor must accept that proof and then authorize a payout. This dynamic itself has one unfortunate flaw: it gives the insurer some degree of discretion even though the insurer has an incentive to reject claims.
A smart contract-based system could reduce that discretion. Such a system could automate payouts based on oracles for off chain events. Such oracles could include:
- Seismic monitoring stations: payouts could be triggered in the event that a large earthquake was detected in the policyholder’s area.
- Internet connected devices: Google’s Nest could be used to detect home invasions or fires. Sensors in many modern cars could be used to detect the failure of major components or an automobile accident.
- National weather service: droughts, floods and other major weather events could be used to trigger payout conditions.
Publication of such events to an immutable and auditable ledger–as proposed above–could further increase the efficiency of smart contracts.
While such a system would restrict some in-the-moment discretion, it does push immense trust onto the oracles themselves, which represent a potential attack vector. In the case of Aigang, Apple is likely to secure its batteries’ APIs. This feature makes distributed insurance attractive for internet connected devices which might have a network owner to secure data publication e.g. this device experienced a failure or this network detected a seismic event. However, not every entity tasked with attesting to claim-triggering events will have that capability. Moreover, not every IOT device will have such security.
Distributed Risk Assessment
One core component to large-scale insurance provision is the ability to pool risk effectively. At heart, this grows out of a more fundamental problem in providing insurance: adverse selection. People who are most in need of insurance are more likely to purchase it, but policy providers are not necessarily in the best position to know this.
Many modern insurance providers address this by collecting as much information as possible to build risk profiles and price policies accordingly. Policies vary over time, by customer, by event, by geography–by every variable perceived to impact price. Such models are part of insurance provider’s secret sauce and why they employ legions of actuaries. A blockchain doesn’t necessarily obviate the need for this or render the current approach irrelevant: a centralized insurance provider could still collect on-chain and off-chain information to sell smart-contracts with baked-in pricing.
This can be further distributed, and many blockchain solutions instead rely on a prediction market for such pricing. Market participants can bet on whether an event is likely to occur: ‘premiums’ can be as simple as placing small bets that an event will happen. They have the added benefit of augmenting oracles for claims verification: market participants have a vested interest in the prediction market resolving truthfully regardless of who buys or sells the insurance products.
However, prediction markets also rely on liquidity and benefit immensely from network effects. Launching specialized prediction markets is no mean feat.
Additionally, existing literature notes several practical challenges to a blockchain-based insurance system. These include network scaling considerations for widespread adoption, the legal interoperability between existing insurance contracts and smart contracts, and UI/UX design to bolster customer understanding of insurance policies. While these are legitimate considerations, they are not convincing arguments against the use case in general and are somewhat premature given the general state of distributed insurance.
The opportunity and disruptive potential that blockchain technology has for the insurance industry has not gone unnoticed by leaders in the business community. Deloitte and Accenture, major consulting and auditing firms, both have teams working on applications of blockchain technology in insurance. In 2016 PwC, a big four auditing firm based in London, released a proof of concept for a blockchain based ecosystem designed for the insurance industry. In Q2 2017, American insurance giant AIG went a step further and announced a partnership with Standard Charter to deliver a smart contract powered multinational insurance policy that uses Hyperledger Fabric as its base blockchain.
Etherisc, a project mentored by CoinFund’s Jake Brukhman, is another attempt to launch distributed insurance markets. The economic model is more sophisticated than Aigang’s and involves distributed collateralization of policies, risk pools backed by reinsurance pools, and automated claims processing. They released the first prototype for their flight delay insurance product in September of 2016. Later that year they would go on to win a number of blockchain design competitions including 1st place at the Blockchain Startup Contest in Graz and as most funded most funded project in hack.ether.camp, securing a total of about 50,000 USD in funding. In the last year, Etherisc has launched running experiments for automated crop insurance and social insurance on Ethereum. They plan on launching a token generation event with Cofound.it some time in 2018.
Aigang in Context
Ultimately, Aigang proposes leveraging all three of these technology applications.
Immutable auditable data: Aigang will partner with manufacturers who can attest to data integrity and serve as oracles. This approach should yield trustworthy data, though it limits the ultimate application of Aigang’s insurance products to manufactured devices with a network operator. This also makes Aigang’s insurance product more akin to a warranty.
Automated claims processing: Aigang will rely on Ethereum smart contracts for claim processing. Aigang doesn’t address issues of either smart contract failure or any policy re-negotiation, though the fully automated nature of its claims processing could render the importance of customer service less relevant.
Distributed risk assessment: Aigang will attempt build its own prediction market to insure these devices. This is likely the weakest link in Aigang’s plan, for several reasons. First, a public market is less well positioned to price things like device performance (manufacturers have the best data), and such markets would suffer from complex challenges that Aigang hasn’t yet thought through (manufacturers are akin to inside traders). Second, building a custom prediction market is not easy and will take time. The fact that the token being sold is intimately tied to the prediction market makes it all the more challenging. They will need to track and implement any innovations in distributed prediction markets that emerge from the industry.