URBAN OS: connecting property development through digital twins and AI

Никита Ивановский — Fri, 13 Nov 2026 18:10:51 +0000

By Nikita Ivanovskiy, founder of IVCORP / URBAN OS

English editorial adaptation of my Russian article prepared for NOPRIZ. This describes our approach and development programme; it is not an endorsement by NOPRIZ. The English adaptation was prepared with AI assistance.

Cover: AI-generated conceptual diagram of URBAN OS, a residential digital twin and twelve development stages. Illustration from IVCORP materials; not a photograph of a delivered project.

Property development brings together investors, architects, engineers, contractors, sales teams, operators and residents. They work on the same asset at different times, yet the reasons behind their decisions often disappear between handovers. A drawing survives; the explanation of why it was approved may not. A building model survives; its connection to installed equipment may not.

At IVCORP, we are developing URBAN OS around continuity: connecting objects, documents, versions, responsibilities and services throughout a development project's life. Our ambition is to automate repeatable work across twelve lifecycle stages while retaining accountable professional review.

A digital twin needs a maintained relationship with the asset

An architectural render communicates appearance. An interactive 3D model lets a user explore space, compare options and retrieve prepared information. A digital twin adds a maintained relationship with the state and lifecycle of the real asset. These tools can share geometry, but they serve different purposes.

For a sales team, useful data may include an approved layout, the available fit-out and a unit's current status. For an operator, it may include equipment identifiers, maintenance records and the documents supporting an intervention. Update frequency depends on the task; every object does not need a live sensor.

Select a small number of elements, give them stable identifiers and connect them to documents, owners and status changes. Test whether a person working with the asset can answer a real question more reliably. Expand the scope after the workflow works.

An AI core should organise context and tools

Our AI approach combines language and multimodal models with a knowledge base, retrieval, specialised tools and controlled agent workflows. An assistant receives a task, accesses the data allowed for its role, prepares an output and passes consequential decisions to the responsible specialist.

Retrieval and fine-tuning solve different problems. Retrieval supplies current project information and its sources. Fine-tuning may help with recurring structures, terminology and output conventions. Neither removes the need to evaluate answers on material not used to build the workflow.

Geometry, calculations and engineering checks should run in tools where their results can be reproduced. An AI-generated explanation helps when a professional can inspect the basis of a decision. Fluency alone is not evidence of correctness.

One development goal is to shorten preparation of design and working documentation. The ambitious time horizons discussed in the original article are targets for future pilots, not measured delivery commitments. They depend on prepared inputs, a defined output scope, suitable tools and professional validation. External approvals, regulatory review and construction schedules remain separate processes.

Twelve stages need explicit handovers

The broader lifecycle connects the investment idea, site and constraints, concept, design, documentation, construction, marketing and sales, fit-out and personalisation, handover, operations, community services and future renewal. Organisations may group these activities differently. The useful question is what the next participant actually needs.

A handover should preserve identifiers, approved versions, change history, evidence and responsibility. It should distinguish public material from operational or confidential information. A shared platform does not mean every participant sees every document.

Start with one transition, such as design to operations. Choose several equipment elements and ask the receiving team to complete a task using the handover. Missing identifiers and unclear responsibilities become visible before the process is scaled to a portfolio.

From 3D visualisation to a connected customer journey

Architectural visualisation, exterior and interior renders, animation and digital locations can give buyers a clear view of an unbuilt project. A website, touch table and immersive sales room need different interfaces and levels of detail while keeping approved information consistent.

A connected experience can preserve a buyer's chosen unit, fit-out and questions when they move from a website to a meeting. The transfer should be understandable and based on consent. A concept image must remain distinguishable from an available product option.

After a sale, the user needs different services: instructions, reliable building information and a way to request support. We treat the sales experience as part of the asset lifecycle. Commercial benefits should be evaluated through a defined pilot, with comparable audiences and explicit metrics, rather than assumed from engagement alone.

Education, community and a digital cities centre

The IVCORP Academy, professional community and proposed Centre for Digital Cities and Virtual Development complement this programme. We want learning to connect to practical tasks: preparing a brief, inspecting a model, developing a service or evaluating an AI-assisted workflow.

The Centre is a developing concept, including education, research, product studios and demonstrations. Its architectural images depict that concept. Physical implementation requires partnerships, a site, design work and delivery. An online demonstration offers a way to explore the idea with potential partners.

Our research interests include urban simulation and the relationship between environments and human experience. Long-term scenarios are hypotheses to examine with explicit assumptions and uncertainty, not validated predictions of individual behaviour or city development.

About the founder and the current project

I founded and developed PLAYESTATE and subsequently left the company. My current work is IVCORP / URBAN OS, a new team and a broader programme of development automation. Historical work retains its original attribution; it should not be presented as a newly delivered IVCORP project.

We welcome specific pilot proposals from developers, architecture and engineering teams, universities, regions and technology partners. A useful starting point is one asset, a defined problem, agreed inputs, responsible people and criteria for judging the result.

Original illustrated article: https://urban-os.online/blog/nopriz/
Ecosystem presentation: https://urban-os.online/ecosystem/pres/
Company and founder: https://ivcorp.online/company/nikita-ivanovskiy/
Contact Nikita: https://t.me/naiv3d