When our team started the project SeniorWard, our vision was simple: seniors are victims of increasing scams, and we want to help them. This is a problem worldwide, and it is difficult to solve. However, in the end, I believe we found a promising solution: AI-powered detection based on both conversation data and transaction data, and compensation for their loss. Here I will review the three major changes that happened during the project.
Compensate the customer
The compensation mechanism is the idea I proposed. I am fond of it and believe it makes SeniorWard stand out from similar anti-scam products.
It originates from the monetisation problem that arose in the early stages of the project. In the very beginning we were aware that we should choose a venture for the project, as our vision required the integration of multiple sources of data, and an existing company would not be so flexible. And then came the monetisation problem, a problem every venture must face. Similar services are either public services or part of the risk management system of big organisations (for example, Visa or Mastercard), so even if customers are not willing to pay, the services will still be sufficiently funded. This is not the case for SeniorWard. It needs to be profitable to maintain itself, but there is no evidence of customers’ willingness to pay.
The team came up with the idea of positioning SeniorWard as a public service, so that it would be likely to receive governmental funding. Just before we made that final decision, some thoughts came to my mind: people are reluctant to pay because they (at least I) can recognise the scams most of the time. Even so, many seniors continue to talk with the scammers, partly because of loneliness. Only when they lose money will they feel regretful. So the problem is the loss of money, and if we can solve this problem, we can charge a high price. These thoughts later became an insurance-like business model of SeniorWard: we charge higher prices, because we compensate for their loss.
From vision to prototype
During the development of the prototype, we ran into exactly the problem of data integration again, but from a different perspective: how can a start-up cooperate with all the necessary parties? In the proposal stage, my idea was that SeniorWard would act as an insurance company and get data from social media providers. However, later research showed that scams happen across social media, phone calls, messaging apps and many other channels, and seniors may use many different banks. Therefore, cooperating with all parties means long-lasting negotiations, which is unaffordable for a start-up.
So we discussed it in the group chat, and I mentioned a two-stage idea: in the beginning stage, SeniorWard only focuses on the most concentrated scam channel; only when the business model is proved will SeniorWard expand to other channels (the expansion stage). This idea was welcomed by the team and later turned into a prototype. In the prototype, SeniorWard will only focus on one bank, one channel (the phone call), and will use a physical device attached to the back of the phone, so that the device can listen to the phone call directly and bypass the data-access problem entirely. Also, the insurance compensation is no longer an in-house function but is provided by an insurance company as a partner.
Re-think about our relation with the bank
In the final stage of the project, I was responsible for the ecosystem and competitor analysis, and also for the evaluation. The more research I did, the more complicated the relationship I found between SeniorWard and the bank. First of all, SeniorWard’s service is built on the bank’s transaction data, so the bank is our supplier. And due to a new EU regulation stating that banks should pay customers for losses caused by fraud from 2027, I believe that the bank will be a good partner, or the so-called “complementor”: a stronger SeniorWard leads to fewer losses for the bank, and thus to deeper cooperation, which can build an even stronger SeniorWard. However, I later found that some banks are also trying to build their own anti-scam systems, which shows the possibility that the bank may instead be a competitor, or an enveloper, since they have greater bargaining power and access to the transaction data. This made me genuinely unsure how to position the relationship with the bank.
My teammate proposed a cleaner solution: SeniorWard would charge the bank directly, rather than the individual. I find this convincing, because it addresses two problems at once. The monetisation problem becomes easier, since earning revenue from millions of individuals is far harder than earning the same amount from a single B-end customer that you don’t need much effort to persuade (remember the new EU regulation). The defensibility problem improves as well. Once the bank is our customer, our relation with it becomes direct, and the envelopment threat, which largely comes from the money issue, turns into an argument we can negotiate.
In retrospect, however, I would say this also makes the potential conflict more explicit in terms of money. And it is conceivable that SeniorWard must reach a higher accuracy rate, because the bank, as our customer, would explicitly demand it. So this solution is not a solution with only advantages.
Final summary
If I had to name the most surprising thing about this project, it is the fact that the major changes were driven by evidence rather than by theories I learnt in the courses. Theoretical framework can help you think systematically, or in a MECE way as some consultant guys may call, and probably offers some insights; but it can hardly provide new information and lead you to a decision.
The second surprise came from collaboration itself. It’s cliché that people should discuss and be open-minded. But in reality, it would be better to divide the task into different parts and everyone works on his/her own part. This makes the responsibility clear, and is generally more effecient. However, the project really taught me that a team can divide the task, but they still need to discuss and share ideas. A problem that keeps confusing you may be easily solved by a teammate’s muse.
