During recent storms in New York City, small businesses have been forced into repeated cycles of flooding and recovery. In Sunnyside, Queens, the owner of Finest Sushi has experienced flooding more than ten times in the past two years, with water surging up through basement drains and reaching depths of over two feet. Each event forces her to halt operations, replace supplies, and spend hours cleaning and preventing mold, costing approximately $7,000 and more than 80 hours in recovery efforts. Despite these repeated incidents, her business is not located within an official flood risk zone, limiting access to city support programs and leaving her to manage the impacts largely on her own
This pattern is becoming increasingly common across the five boroughs. When high-intensity rainfall overwhelms the city’s drainage systems, the financial burden falls disproportionately on ground-floor and basement businesses. Losses quickly accumulate through damaged inventory, equipment failure, and forced closures. At a broader scale, the September 2023 flooding event resulted in an estimated $100 million in total economic losses across the New York metro region. For an independent restaurant or neighborhood retail shop, a single severe weather event can erase months of profit and threaten the viability of the enterprise.
The root of this imbalance is not only physical infrastructure, but informational infrastructure. Tools produced by the Federal Emergency Management Agency and other agencies are not designed to capture rainfall-driven flooding at the block or storefront level. As a result, many businesses operate in what is effectively a blind spot, where risk is experienced but not formally recognised.
Public investment has largely focused on long-term physical solutions. Programs led by the New York City Department of Environmental Protection (DEP), including the Cloudburst Program and sewer system upgrades, aim to reduce citywide vulnerability. But these interventions take years to implement, and when water is already entering their space, a small business wonders is often stuck somewhere between a water-prone building they do not own (owned by landlord) and waiting for the city to improve sewers (and only IF they are in an area prioritised for expansion).
This gap has created a different kind of market failure. Information about flood risk, response actions, and recovery is fragmented, inconsistent, and often inaccessible at the moment it is needed most. Small businesses are left to improvise under pressure, making decisions with limited guidance and no reliable way to learn from previous events.
FloodLine is built to bridge this gap. Developed through programs at the Massachusetts Institute of Technology, including the Delta V accelerator, and currently participating in initiatives at the Cambridge Innovation Center (CIC), the platform focuses on improving how flood-related information is structured and delivered at the local level.
Instead of introducing new physical infrastructure, FloodLine focuses on improving how information is accessed and used at the moment of risk. The platform is built as a mobile-first system designed specifically for small businesses operating at street level, where flooding impacts are most immediate.
At its core, FloodLine provides a set of tools that support small businesses before, during, and after flood events. This includes weather notifications ahead of severe rainfall events based on localised weather conditions, allowing them to take early preventative action. During a storm, businesses can find mitigation tips, nearby local resources to help as well as document conditions through photos and short reports, creating time-stamped records of water entry, damage, and response actions. After the event, this documentation remains accessible, helping Business Improvement Districts (BID’s) track patterns over time and organise recovery efforts.
The platform also connects users to relevant resources, including local emergency contacts, cleanup services, and guidance tailored to their specific situation and location. Rather than requiring users to search across fragmented sources, FloodLine centralises this information into a single, accessible interface designed for use under pressure.
A key component of the system is FloodChat, a multilingual assistant that delivers real-time, location-specific guidance. Developed as a retrieval-based model, FloodChat structures and surfaces information from curated datasets to provide clear, actionable responses during flood events. The system is designed to prioritise usability and speed, ensuring that businesses can quickly access relevant information curated from high-quality sources without losing time (and money) searching and finding and reading through often repetitive information. Also making it available in 35 languages which is critical in a diverse urban population.
This work, building on a foundation first funded by a Rebuild By Design grant, is now led by Maysaa Sati, who focuses on refining how information is organised, retrieved, and delivered within the platform. Her work centers on improving the system’s ability to translate fragmented, real-world inputs into usable guidance that supports decision-making.
Beyond individual users, FloodLine is designed to support Business Improvement Districts (BIDs) and other business-serving organisations. By aggregating anonymised, time-stamped reports across storefronts, the platform helps translate on-the-ground conditions into coordinated, actionable responses at the corridor level. Rather than simply increasing visibility into flooding, it enables local organisations to guide businesses toward immediate mitigation and recovery steps while also identifying patterns that can inform longer-term planning.
“Climate resilience will not improve through bigger datasets alone,” Sati noted. “It requires restructuring how data is collected, translated, and delivered at the ground level.”
The significance of this approach lies in how flood response is being reframed through tools like FloodChat. Rather than treating each flood event as isolated, the system is designed to interpret recurring, place-based conditions by connecting user questions to relevant, localised information. This introduces both challenges and opportunities in designing systems that can respond to highly context-specific needs, where the same question may require different answers depending on location, building type, or stage of the flood event. By structuring information around these variables, the platform begins to move beyond generic guidance toward more precise, situational support.
Over time, this creates a different kind of visibility. Patterns begin to emerge at the scale of individual blocks and storefronts, revealing gaps that are not captured in traditional models. This information has implications not only for individual businesses but also for Business Improvement Districts and local organisations that coordinate response efforts across commercial corridors.
FloodLine does not eliminate flood risk. But it changes how that risk is navigated and assists in the small business journey from risk to mitigation solutions.
As climate events become more frequent, the question is no longer whether cities will flood, but how well businesses are equipped to respond when they do. In New York, where thousands of small businesses operate at street level, that question is increasingly tied to access to information, not just infrastructure.
