Vectoryx positions itself as a regulated multi-party infrastructure for claims processing and, among other things, uses ChatGPT integration to bring the initial claim report directly into the dialogue with those affected. Co-founder Alexander Stade discusses media discontinuities, interoperability, and why claims management rarely fails due to regulatory issues.
Property damage claims are complex events involving many parties: owners, residents, property management companies, brokers, insurers, and restoration partners. This is precisely where Vectoryx comes in with “Claims,” a claims settlement product that connects all parties in real time. With its own app in the ChatGPT Store, the company is also taking an unconventional approach: turning the claims filing process into a dialogue-based AI experience. In an interview withthebrokernews , co-founder Alexander Stadeexplains where the actual pain points in the claims process lie, how regulated infrastructure and AI interfaces interact, and where Vectoryx aims to go as a company.
What was the specific moment or experience that led to the founding of Vectoryx?
Vectoryx was born out of my own personal experience. Over several years, I built up my own property management business and later sold it. That’s exactly where the pain point arose: claims management is an extremely complex process involving a great many parties. And even after consolidating all the information, created a clean database, and submitted it to the insurer, it still took another two to ten months to receive approval to begin the work which severely undermined the satisfaction of everyone involved and ultimately actively “cost” all parties customers.
So, on the one hand, we faced enormous communication challenges when claims arose; on the other hand, we faced enormous coordination challenges such as bringing appraisers or restoration contractors together with those affected, and then figuring out how to get back on track with the process afterward. One of the co-founders comes from the insurance brokerage sector and faced exactly the same coordination and communication challenges with claims involving single-family homes as we did in the housing industry. Added to this is the economic reality: Handling claims is labor- and IT-intensive, but is generally not compensated separately; instead, customers view it as an included service. High effort, no return. That’s how the company came to be. We mapped out the entire process, spoke with damage restoration contractors, insurers, and other property management companies, and used that input to develop our first product, which has evolved significantly since then.
You say that claims management rarely fails because of the claims settlement process itself, but rather because of a lack of information or unsystematic information. Can you illustrate this with a typical real-world example?
Let’s take water damage as an example. At a very high level, this is how it works: The damage is reported to the property manager, who documents it and then coordinates the response. He brings the damage restoration contractor together with the affected parties, informs real estate agents and insurers, and thus manages precisely these communication and coordination efforts. At the same time, they collect data: root cause analyses, leak detection reports, measurement logs, drying service quotes, demolition and restoration quotes, and information on the parties involved which, due to data protection regulations, usually cannot be easily stored proactively, for example, in a condominium tenancy agreement.
This results in a dataset that is transmitted to the insurer. The problem is that, in practice, the claim report usually does not reach the insurer with all the necessary data included. The insurer must review the claim, reconcile the data, and then follow up later because master data is missing, information about the parties involved is missing, the leak location is missing, the quotes are inconsistent, or other details are missing. This delays the entire process, specifically on the insurer’s end. That’s exactly where we come in: We provide the insurer with the data in a structured format and with standardized data quality, so that they can make decisions in minutes rather than months, because all the information relevant to the decision is available.
How has the concept behind Vectoryx evolved since its founding, and what have you learned along the way that you wouldn’t have expected at the beginning?
That, at its core, this is not purely a software problem, but a translation problem. In the insurance industry, we have very long sales cycles. At the same time, appraisers and restoration contractors still perform much of their work manually. So we’re seeing completely different sectors of the economy coming together: the trades through the restoration specialists, the assessment of damage by the appraiser, the insurer with its traditionally highly regulated processes, and the property manager with his or her ever-increasing digitalization. Each of these stakeholders speaks its own language, and we need to bring them all to the table.
A good example is the skilled trades: They are extremely difficult to digitize, and others have already tried to do so. This usually fails not so much because of the technology, but because there is simply no need for it among tradespeople. We have addressed and systematically solved precisely this problem by bringing all stakeholders together in a single process without creating additional hurdles.
Vectoryx connects insurers, property managers, and renovation partners in a single system. Where exactly do the biggest information gaps occur today between these parties?
Basically, wherever there’s a system boundary. And these system boundaries are very diverse. The skilled trades operate differently, including on the software side, than the housing industry. The housing industry operates differently than the insurance sector. Wherever there’s a system gap, these very losses, and sometimes massive efforts to bridge them arise. But this applies just as much to the point at which the people affected are involved. Some are highly tech-savvy, while others for example, an 80-year-old, prefer to communicate exclusively by phone. So we’re dealing not only with system-related gaps but also with generational ones. In this regard, we also act in part as an impact startup, as we actively promote inclusion to the best of our ability and integrate all stakeholders as seamlessly as possible.
You rely on BiPRO and GDV compliance in Germany and their equivalents in Austria and Switzerland as a foundation. How important is regulatory interoperability for acceptance among established insurers, and in what ways does it tend to hinder innovation?
Our processes are modeled after BiPRO and the GDV standard, but we have deliberately chosen not to be fully bound by them. Instead, we focus on ensuring that these various standards are actually implemented in practice. This also includes FRIDA, the insurance industry initiative that aims to use a digital wallet to consolidate this very flow of information among the various parties involved. And, of course, this creates a point of friction, because the question arises as to what fundamental interest an insurer has in standardization. It might even make the insurer somewhat interchangeable as a result: If all information is based on the same system architecture, it becomes more difficult to demonstrate its own expertise. However, that is more of a question for the future.
For us, it really doesn’t matter which system the insurer uses in the first step, because we can integrate with any existing system environment. We can access the insurer’s system directly via APIs. However, we can also use agents to access the insurer’s systems at no additional cost and with very little implementation effort. That’s exactly what we like to do in pilot projects: connect directly to the system and provide information without the need for complex system integration upfront.
With the ChatGPT integration, you’re taking an unusual approach to initial reporting. Why use a consumer interface like ChatGPT instead of your own app or portal?
Of course, we also offer affected parties the option to access our claims process directly via a link or to be notified simply via traditional email. But we have to face reality: There are an insane number of apps out there today. The insurer has its app, the property management company has its app, and in some cases, even contractors now have their own apps. For someone dealing with a claim, it’s then almost impossible to keep track of where the case stands, what the next steps are, and what information is needed.
We could launch the next app right now and say, “Hey, sign up here to see the progress.” We’re deliberately choosing not to do that. Instead, it’s great to integrate into the systems people are already using. The user numbers for ChatGPT and similar systems show that these interfaces have caught on. That’s exactly what GPT integration is for: enabling users to file claims, update information, and check status updates right where they’re already working. And it allows us to do something else: claims can be reported to us even if there isn’t yet a contract with an insurer, a property management company, or a broker. At this level, we’re expanding in all directions.
How do you ensure that a dialog-based, AI-powered claims reporting system actually provides complete and reliable data, especially in cases of complex property damage?
By bringing all stakeholders together. We know exactly what information the various insurers need because we maintain the necessary partnerships. We have databases and integrations for example, to review quotes, and we involve appraisers. This allows us to respond immediately upon receipt of a claim, and that response involves directly engaging restoration contractors, appraisers, brokers, and other relevant parties. In this way, we compile the information right in the claim report so that it can then be transferred to the insurer easily and completely.
A native integration based on the OpenAI Apps SDK is in the works. How will this differ from the current GPT Store model?
We’ve actually already received approval for this, and the launch is coming up very soon. The difference is that users can incorporate our app directly into their daily workflows. They can access our integration from within an existing chat and initiate the claim right there. And they can use it in context: If someone is already chatting with ChatGPT for example, about rent reductions, standard processing times, or questions regarding utility bills, they can specify their specific claim and use our SDK to ask questions directly about it. It’s also becoming interesting from the perspective of the housing industry, since many of the applications there work via GPT interfaces: So if a property manager “talks” to GPT about a property “tomorrow,” the transition to our platform can happen right there, without having to switch systems at all. These store-based transitions may therefore also be possible in the future.
It’s all included in the final release, and there will be quite a bit more added over the next few months.
What does your business model look like? Who pays for Vectoryx, and how does this change as more and more process steps are automated?
Our model is funded, in part, by insurers, since they stand to benefit the most. Insurers are suddenly able to settle claims within minutes rather than months. At this level, we offer insurers the opportunity to save around one thousand euros in internal staffing costs per claim. And, more importantly, they increase policyholder satisfaction, thereby setting themselves apart in the market.
Automated, structured processing, and, above all, staying on top of things for the customer and proactively keeping them informed, is entirely our responsibility. So we don’t just react, we actively provide information. The key point is: We don’t deliver software that needs to be set up and maintained by someone else, as was the case with traditional SaaS models. We deliver a solution. And this solution benefits the insurer the most, because they derive the greatest value from it. Looking to the future, AI technologies will naturally continue to streamline processes, and as an AI-native, agent-based company, we’re strongly committed to this. However, it’s also important to note that every single insurance claim can be so complex and open up so many paths left, right, or straight through the middle, that we see ourselves more as a deep-tech company than as a traditional insurtech company.
Where do you see the greatest resistance from insurers and administrative bodies when it comes to implementing a cross-platform claims infrastructure?
For insurers, the main source of resistance is that they still view many processes such as claims intake as something they must handle themselves. So a shift in mindset and an openness to AI-native solutions are needed to improve their own service quality. This is a traditional issue in the industry that has developed over time: Insurers are large companies with lengthy decision-making processes, and their cycles are correspondingly long. It simply takes time to make inroads there.
However, we have some very exciting pilot projects in the pipeline that will allow us to handle corresponding volumes. On the other hand, we see virtually no resistance from property management companies, since our service is free of charge for property managers. We reduce their workload per claim by an average of about ten hours, spread over several months. We provide our property management clients with a dashboard, and we have clients for whom we’ve already saved hundreds of workdays in staff and system resources, time that can now be put to other uses.
By the way, our clients include not only traditional third-party property managers but also property owners and investors. And since we operate as a white-label solution, no one realizes that we’re working behind the scenes. We improve service quality and customer satisfaction on behalf of our clients. Consequently, there are no negative concerns about our services within the housing industry.
You are currently operating primarily in the DACH region. How scalable is your model across national regulatory boundaries?
We’re already active in the U.S. and have our own branch there. The fact is, “insurance claims” exist in every country in the world. There’s always a home insurance provider, and ultimately, there’s someone who has to repair the damage. The process we’re establishing here can therefore be replicated in any country certainly with one or two regulatory adjustments, but in principle, it’s always the same. This works in Germany, France, the UK, the U.S., and also in Asian markets. Accordingly, we’re currently planning a major expansion into other countries.
How much discretion do you give AI in the claims process, and where do you deliberately draw the line when it comes to human review?
As a rule, we do not make a settlement decision right away. Once a claim has gone through our quality pipeline which includes, among other things, verifying the cause, the insurance policy, and the terms of coverage we know with a very high degree of certainty whether the claim is covered or not. However, we do not settle the claim at this stage; the decision remains with the insurer.
However, we collaborate with certain insurers to the extent that we receive claims approval authority: If claims meet our quality criteria, we can approve high-frequency claims up to a defined approval limit from the outset, so that the insurer does not incur any additional effort afterward. In this way, we handle high-frequency claims, which in some cases account for up to sixty percent of all claims.
An important point in this context is the “zero-claim” report. When a claim first arises, no one knows whether it is even covered by insurance; that only becomes clear as the process unfolds and after the cause has been analyzed. Consequently, we process a large number of claims that ultimately turn out not to be covered. This helps us shed light on the “hidden claims” that currently exist among insurers: claims that aren’t covered, claims below the deductible that are never reported, or claims where a broker or administrator actively decides not to report them to avoid increasing their own loss ratio. The claim is effectively absorbed. The result is that accurate underwriting and risk calculation are actually impossible.
That’s exactly what we resolve, because we’re based at the property itself. It doesn’t matter to us whether there has been a change in insurer, broker, property manager, owner, or tenant in the past (or in the future), because we’re always based at the property. This enables us to provide historical master data and historical risk data, something no one else has been able to do until now, and to eliminate the blind spot in underwriting.
Ultimately, however, we also have employees who actively step in during more complex situations to clarify issues or make decisions.
What role does trust play for both policyholders and insurers when an AI system becomes the central interface for claims?
It’s important to understand that, as a rule, we initially only provide guidance. The actual decisions are made by the various authorities. What’s crucial is that we keep people engaged throughout the entire process. After all, in most cases, it’s not that someone necessarily wants to speak with a person. When an affected party calls the insurer, the restructuring firm, or the administrator, they usually want information because their need for information hasn’t been met. That’s exactly what we manage from the outset by keeping affected parties and every other stakeholder so well informed that it’s clear at all times when what is happening. The major advantage is that there are no longer any blind spots in these processes. And because these blind spots are eliminated, trust is significantly higher than in traditional claims processing, which requires a large allocation of personnel resources that are, by their very nature, susceptible to vacation, illness, and staffing shortages. We act as a highly reliable partner.
Where will Vectoryx be in three years? Will it remain a claims management tool, or will it evolve into something bigger such as a data infrastructure for the entire real estate and insurance industries?
We see ourselves in the future as the central data interface for insurance claims in the housing sector. Our core capability will always be claims management, specifically, AI-native and agent-based claims management. Of course, there are also opportunities to further develop the product. We have a vast amount of risk data and, as described, can provide reliable risk index data. We also have an enormous amount of data on restoration quotes, prices, and price trends. This enables us to provide comprehensive advice to both the insurance and real estate industries on how risks can be mitigated—for example, through maintenance measures. And this risk mitigation can be factored into insurance premiums. If an insurer lacks risk information, we can provide exactly that.
There’s another important point to consider: Homeowners insurance is currently extremely unattractive to many insurers. This is evident from the fact that many are withdrawing from the market; in each state, there are really only a handful of truly relevant homeowners insurers left. This is because claims are becoming increasingly expensive and the processing time is getting longer and longer. In the United States, the average turnaround time until the insurance company responds is over 55 days, and this figure is rising sharply. In Germany, too, this has been the subject of strong criticism for quite some time. With our service, we are able to restore building insurance to a level that is attractive to insurers and gives property owners, as policyholders, the assurance that their buildings are realistically insured—and, among other things, helps them reduce their own liability.
If you could give just one piece of advice to someone who reports water damage in their apartment tomorrow, what would it be?
Please be understanding of the typical processing times of the various parties involved. Although they usually already work together, they do not operate according to the kind of time standards that are generally expected these days.
At the same time, it must be honestly pointed out today that processing times at the insurer are very long. Neither the administrator nor the parties involved in the trades are to blame for this, these are processing times that affect the entire industry and are partly due to the fact that lawmakers impose requirements whose scope is often difficult to grasp.
We’re trying to clarify the situation here: These delays do exist in the market, and they place a burden on those affected. But times are changing, and there are certainly ways such as through our services to handle claims more effectively and significantly faster.
How do you handle data protection when an AI system aggregates claims data across multiple parties?
We naturally adhere to applicable data protection standards and provide our services exclusively within the framework of commissioned data processing. This means that the data subject, as well as all other stakeholders, are involved in such a way that each party receives from us only the data that is strictly necessary to fulfill the legitimate interests of the parties involved. Our approach is to store data only to the extent that it is actually necessary and, where appropriate, even to anonymize it during individual process steps. We expressly consider these standards to be sensible. Here, too, we provide the insurer with another major benefit: For example, we pre-filter data during the reporting process so that it poses no data protection risks. The details are where things get interesting, as there are indeed significant differences between, for example, the German and American markets, but that is a topic for another time.
Binci Heeb asked the questions.
Alexander Stade is the founder and CEO of Vectoryx. Before founding the company in 2025, he spent several years building his own property management firm specializing in condominium and rental property management. The inefficiencies he encountered in claims management during that time served as the inspiration for Vectoryx.
Through Vectoryx, he is currently developing an AI-native infrastructure that brings together insurers, property managers, insurance brokers, and renovation partners in a single process. The company is based in Hamburg and New York; in 2026, Stade and Vectoryx were selected for the Founders Program at the Global Insurance Accelerator in Des Moines.
See also: Swiss InsurTech Hub Looks Back on an Eventful First Half of the Year
Search:
Sponsors:
