The most honest thing about a city is its streets. They are a messy, multimodal ledger of everything that happens there, from a cracked curb to a faded crosswalk to the particulate count in the air. Most of that data is either uncollected or locked in siloed, unqueryable formats. Ensense AI, a small team in Culver City, is betting that the first company to build a coherent, queryable data layer for the physical world will own a fundamental piece of the urban stack. They are not raising money to do it, at least not publicly. They are just building.
The bet on a queryable street
The core of Ensense is a proposition that sounds simple but is technically vast. The company collects and indexes multimodal street-level data,video, imagery, and other sensor feeds,about infrastructure, traffic, signage, and environmental conditions [YouTube, June 2024]. This raw data lake is then exposed through an AI question-answering layer they call Corvect, which allows users to ask complex spatial and temporal questions in plain language [YouTube, June 2024]. The vision is to move from measurement to intelligence in one package, a data cloud for streets and cities [Ensense AI, Unknown]. For a city planner, that could mean asking, "Show me all intersections with faded pedestrian markings installed before 2018." For a utility, it might be, "Identify all manhole covers within 50 meters of trees showing root uplift." The unit of value is the answered question, not the terabyte of raw sensor feed.
The founder's energy-to-infrastructure pivot
Leading the technical charge is Shahram Farhadi, identified as both CTO and CEO of Ensense [The Org, Unknown] [YouTube, June 2024]. His background is not in urban planning, but in applying AI to complex physical systems. He holds a Ph.D. in petroleum engineering and previously served as Head of Energy Technology at AI company Beyond Limits, developing solutions for the energy industry [Hart Energy, 2026] [Kurdistan Onwards Conference, 2026]. This is a relevant pedigree. The oil and gas sector has long dealt with the challenge of making sense of vast, unstructured sensor data from remote physical assets. Farhadi's pivot suggests he sees a similar pattern in city infrastructure: a fragmented, sensor-rich environment starving for a unified intelligence layer. The team remains lean, listed at 2-10 employees [LinkedIn, Unknown], focused on proving the core technology before scaling go-to-market.
The market of unanswered questions
The potential customers are clear, even if named deals are not yet public. Ensense targets both the enterprise and public sectors [Craft.co, Unknown]. Use cases sketched out include infrastructure management, environmental monitoring, hazard detection, and walkability analysis [Craft.co, Unknown]. The tailwinds are real. Municipalities are under pressure to modernize, manage assets proactively, and justify infrastructure spending with data. Insurance companies, logistics firms, and real estate developers all need better ground-truth data about the physical environments where they operate. The platform's promise is to turn the city from a black box into a searchable database.
Where the rubber meets the road
The ambition is compelling, but the path is paved with hard, non-AI problems. The primary challenges are not in the query layer, but in everything that comes before it.
- Data acquisition and cost. Building a comprehensive, fresh street-level dataset requires either a massive fleet of owned sensors or expensive partnerships. The capital and operational intensity of data collection is the traditional graveyard of similar ambitions.
- Privacy and regulation. Continuously collecting multimodal video and imagery data across public spaces is a regulatory minefield. Navigating municipal data ordinances and public privacy concerns is a full-time operational and legal burden.
- The integration slog. For the data to be actionable, it must plug into existing city and enterprise workstreams,maintenance ticketing systems, capital planning software, compliance dashboards. This is a sales and engineering grind far removed from the elegance of an AI Q&A demo.
For Ensense, the lack of public funding is a double-edged sword. It signals a focus on product depth over hype, but it also raises the question of runway for tackling the capital-intensive data acquisition phase. The company's success likely hinges on a razor-focused initial wedge: picking one high-value, tractable data type or geographic region to master before expanding the multimodal promise.
Back of the envelope, if a single city inspector saves 30 minutes a day by querying Ensense instead of manually reviewing disparate reports, that's about 125 hours of reclaimed productivity per year. Scale that across a department, and the efficiency math starts to work, even before you factor in avoided costs from catching a failing pipe or a hazardous sidewalk. The company to beat isn't another AI startup. It's the incumbent workflow of PDF reports, siloed dashboards, and windshield surveys that currently passes for public place intelligence. Ensense is betting that a coherent data layer can finally make that system obsolete.
Sources
- [Ensense AI, Unknown] Homepage | https://ensense.ai/
- [YouTube, June 2024] Ensense: Unlocking a City's Data With AI | https://www.youtube.com/watch?v=pbvzESfD8vE
- [Craft.co, Unknown] Ensense AI Executive Team | https://craft.co/ensense-ai/executives
- [LinkedIn, Unknown] Ensense AI LinkedIn Profile | https://www.linkedin.com/company/ensenseai
- [Hart Energy, 2026] Shahram Farhadi Profile | https://www.hartenergy.com/40-under-forty/Shahram-Farhadi
- [Kurdistan Onwards Conference, 2026] Shahram Farhadi Speaker Profile | https://www.kurdicon.com/shahram-farhadi%E2%80%8B/
- [The Org, Unknown] Ensense AI Company Profile | https://craft.co/ensense-ai/executives