Building Says Hi: An AI Building Brain That Grows Dhaka's Gulshan Market in Three Stages

Q: Who wrote "Building Says Hi: An AI Building Brain That Grows Dhaka's Gulshan Market in Three Stages"?

This article was written by UNI Editorial and published on uni.xyz, a platform for design professionals and creative thinkers.

Q: What is "Building Says Hi: An AI Building Brain That Grows Dhaka's Gulshan Market in Three Stages" about?

This article covers topics related to Architecture of Tomorrow, Advanced Architecture, Human Augmented Intelligence, Smart Buildings, Urban Design, Commercial Buildings. Shariful Anik's shortlisted UnIATA '19 entry gives a Dhaka market a learning brain, letting sensor-laden cubes expand outward, then upward.

Most buildings are finished the day the scaffolding comes down. Building Says Hi argues the opposite: a building should keep listening, learning and growing long after it opens. The project embeds artificial intelligence as a "building brain" that gathers data from its users, simulates the changes coming at it, and adjusts its own form in response. The provocation is not that a structure can hold sensors, since plenty already do, but that the architecture itself should be planned around continuous change.

Designed by Shariful Anik and shortlisted in UnIATA '19, the proposal tests this idea on a demanding site: Gulshan Market in Dhaka, a 4.1-acre commercial ground. Rather than imagining AI architecture on a blank plot, Anik places it inside a dense, informal and heavily used market, where the gap between how a building is designed and how it is actually used is impossible to ignore.

A Story of Preserved Oak Reframes the Architect's Role

Sketch illustration of a group gathered around a table of plans, with a perspective drawing of a window-lined room above

The opening sketch, a group gathered over plans beneath a window-lined room, sets up the project's founding anecdote. Anik cites a 150-year-old wooden structure at Cambridge University that was revitalised using preserved oak reserves, a case where foresight built into the original decision paid off generations later. The lesson he draws is that good architecture anticipates its own future repair and adaptation.

From this story the project extracts three abilities it expects of contemporary buildings: physical sustainability through eco-friendly materials, resilience against unpredictable environmental change, and anticipatory features that prepare for uncertainties such as climate impacts. AI is positioned as the tool that makes the third ability practical, through data mining of user interactions, data processing that simulates and predicts environmental change, and a learning intelligence that synchronises the building with its surroundings.

Gulshan Market's Rooftops, Car Parks and Tin Stalls

Aerial view of a market rooftop with a scaffold billboard frame above a crowded parking lot in hazy daylight

Collage of corrugated tin market stalls, a multistory shopfront block and smoke billowing over a burning market as crowds watch

The site documentation is unsentimental. An aerial view shows the market rooftop topped by a bare scaffold billboard frame, hovering over a crowded parking lot in hazy daylight. A collage follows: corrugated tin stalls, a multistorey shopfront block, and smoke billowing over a burning market while crowds look on. Together these images describe a place where space is consumed by cars, signage and makeshift structures, and where vulnerability is visible.

This is the brief the building brain has to answer. Anik frames the redesign around user-centric insights, specifically consumer behaviour and spatial inefficiencies, which gives the AI layer a concrete job: to read how people actually move through and occupy the market, rather than to optimise an abstract model.

Tarpaulin Streets as Evidence of Unmet Demand

Fabric stalls under red tarpaulins line a narrow street, with sunlight pooling on the pavement and shoppers walking ahead

Paired street views of a market block draped in blue tarps beside buses, and a busy junction with rickshaws and vans

At street level, fabric stalls shelter beneath red tarpaulins along a narrow lane, sunlight pooling on the pavement as shoppers pass. Paired views show a market block draped in blue tarps beside parked buses, and a busy junction choked with rickshaws and vans. The tarps are the most telling detail: they are the users' own adaptive architecture, improvised shade and enclosure added wherever the permanent building falls short.

Read this way, the project's later moves make sense. Shaded areas, semi-outdoor zones and public hubs in the proposal are formalised versions of what traders and shoppers already build for themselves. The design does not import a foreign typology; it learns from the street, which is exactly what it asks its AI to do.

An Orange Plaza Ringed by Semi-Outdoor Cubes

Axonometric rendering of an orange-paved plaza ringed by white gridded volumes, pools and trees, with hot air balloons and birds overhead

The axonometric replaces the car park with an orange-paved plaza enclosed by white gridded volumes, interspersed with pools and trees. Hot air balloons and birds drift overhead, a deliberately optimistic image that contrasts sharply with the haze of the site photographs. The gridded volumes are the project's semi-outdoor cubes, the modular unit that structures everything that follows.

The cubes work as activity zones and shading devices at once, and their open frames suggest permeability rather than enclosure. Their repetition matters: a modular grid is what allows the market to grow incrementally, and it gives the building brain a legible vocabulary of units it can add, adjust or reconfigure as demand shifts.

Archean to Apex: Growing Outward, Then Upward

The project's growth strategy unfolds in three named phases. In the Archean Stage, functional zones are established around "activity magnets" that draw people in. The Augmenting Stage expands horizontally to meet occupant demand. The Apex Stage turns vertical, stacking residential, commercial and community space to support growth while minimising the footprint on the ground.

The geological naming is apt, implying slow accretion rather than a single construction event. It also clarifies the role of the AI: the brain does not design the building in one stroke but decides when and where the next layer should arrive, using what it has learned in the previous phase. For a dense city like Dhaka, the move from horizontal to vertical expansion is the most pragmatic part of the scheme.

Sensors, Screens and Conveyor Robots in the Cubes

The final layer addresses how the building behaves day to day. Each self-sustainable cube carries compact sensors for real-time environmental adjustment, while AI-driven insights shape how the public spaces respond to the people using them. Retail is reimagined too, with interactive screens and conveyor robots proposed for a more seamless shopping experience.

Here the proposal is at its most speculative, and its retail technology sits a little uneasily beside the tarpaulin-and-tin economy documented earlier. The more convincing idea is the sensor-equipped cube itself: a small, repeatable unit that monitors its own conditions is a plausible step toward the responsive market the project describes.

Why This Project Matters

Building Says Hi treats AI not as a gadget bolted onto a finished design but as a planning logic. By tying its building brain to a modular cube, a phased growth strategy and a real site with documented problems, Anik turns a broad claim about intelligent architecture into a sequence of spatial decisions that can be argued with and tested.

Its strongest insight is that Dhaka's markets are already adaptive; traders reshape them daily with tarps and stalls. The project proposes giving that informal intelligence a permanent framework that can expand outward and upward as the city demands. Whether or not conveyor robots ever reach Gulshan, the idea of a building designed to keep learning from its users is one worth carrying forward.