ENGINEERING-LED AI · DUBAI, UAE
We build AI and software that help engineering teams understand changes, test decisions and connect the whole project, from design to site. Based in Dubai. Built around your real problem.
Engineering . Construction . Maintenance . Inspection . Technical Operations
One changed design. A clearer view of what it affects.
01 / THE PROBLEM
We help you see what changed, understand what it affects and decide what happens next.
AI, in plain English
We turn the information your team already has—drawings, models, measurements and documents—into tools that help people make better decisions.
Illustrative possibilities. People remain responsible for engineering decisions.
What we build
From an engineering question to a working tool: we assess the problem, build a focused pilot and test the result with your team.
Understand how a design or site change affects the wider project.
Connect model revisions, quantities, specifications and project dependencies. Agree the engineering checks and validation evidence before deployment.
Explore this service →Explore a digital view of a physical building, asset or operation.
Connect an agreed model with operational data to explore condition, performance and scenarios. The available sensors, data quality and engineering models determine what a twin can reliably show.
Explore this service →Build tools that fit how your engineers and teams actually work.
Custom applications, computer vision, data dashboards and system integrations. Start with a defined scope and measurable acceptance criteria.
Explore this service →Keep everyday work moving while the bigger systems take shape.
Document search, comparisons, reporting, routine approvals and hands-on AI workshops. These can be standalone services or part of a wider engineering system.
Explore this service →We also scope websites and web applications, software integrations, dashboards, and AI-assisted video and audio projects, including explainers, voiceovers and transcription. Feasibility, rights, hosting and delivery requirements are agreed before work begins.
Bring us the problem—even if it does not fit a card.
Engineering & construction / the bigger picture
A design change should come with a clear picture of what it affects. These are the larger systems we want to build with engineering and construction teams.
Concepts and proposed pilots; the demos use fictional data. Open a card to see how each idea works.
Concept · interactive prototype
Connect design, cost, planning, procurement and site around one engineering problem.
Our forward-deployed engineering approach embeds solution development close to your team. The AI Project FDE concept links changes to evidence, affected work and responsible reviewers. It supports decisions; authorised people approve technical, commercial and site actions.
Explore this service →Planned concept
Find what changed in a building model—and what needs checking next.
BIM means building information modelling; IFC is a common model-exchange format. Compare agreed revisions, preserve element references and flag quantity, procurement or programme dependencies for review.
Explore this service →Proposed pilot
Connect the model to the bill of quantities, so quantity changes are easier to review.
A bill of quantities lists the work and materials used to estimate a project. Map agreed model elements to these items, compare quantities and expose missing mappings and unit assumptions. A quantity surveyor checks the result.
Explore this service →Exploration · synthetic illustration
Explore how a digital project view could connect what was designed with what is happening.
A digital twin is a digital representation linked to a real asset through agreed data. SiteLab is our planned exploration of connected project evidence, site observations and scenarios. Live integrations, prediction quality and refresh rates would be established in a scoped pilot.
Explore this service →Try a simple digital-twin scenario
Choose what you want to understand about the same building.
Design: compare an agreed model revision and identify the affected elements.
Illustration only. No live sensors, model parsing or predictive AI.
05 / INTERACTIVE PROTOTYPES
Choose a scenario, change an input and inspect the result. Every export uses the same h.ai PDF report template.
Follow a fictional construction change request from source evidence to engineering, commercial and site handoffs, with human approval points.
Compare sample tender requirements with a quotation, identify gaps and prepare clarification questions for review.
Convert fictional site observations into an inspection report draft and review the proposed actions before release.
Enter task frequency and time assumptions to estimate a workflow opportunity. Results are scenarios, not measured client savings.
These prototypes use calculations and guided rules with fictional data. They show proposed workflows; they are not client results or connected production AI systems. No project approval or message is sent.
H.AI WORKFLOW LAB
DOCUMENT & KNOWLEDGE INTELLIGENCE
Choose a fictional construction file. Ask a question. Open the supporting evidence.
Scripted browser simulation with fictional data. The separate local Windows application is an internal prototype. No confidential uploads, live LLM or client deployment is demonstrated here.
Choose a question, then inspect its cited source. You can export the conversation and selected evidence as a PDF.
About h.ai
h.ai is a Dubai-based AI and software consultancy founded by Haya Alshanqiti. We help teams use models, measurements and project information to solve practical problems.
The engineering question, the people and the available data.
A focused tool, checked against agreed examples and success criteria.
Integrate, train and expand when the evidence supports it.
Construction and engineering are our current focus. Our approach is designed to grow into energy and other industries with complex, real-world problems.
No. Tell us what you are trying to achieve. We assess whether AI, conventional software or a combination is appropriate.
Each card states its status. Our public demos are synthetic illustrations. Production capability, integrations and acceptance criteria are established through a specific project.
Yes. We can scope a focused pilot and assess local or private hosting. Start with sample or redacted information; data access and handling are agreed before implementation.
Meet the founder
Founder & CEO, h.ai Digital Intelligence
I started with the physical world: energy systems, materials and the engineering questions behind them. Data and Python led me into machine learning—and into building tools that help people do something useful with what they know.
The question I keep coming back to: what problem are we actually trying to solve?
Engineering · Applied AI · Technology partnerships
BEng in Sustainable Energy Engineering, Queen Mary University of London.
MSc in Nuclear Decommissioning & Waste Management, University of Birmingham. Distinction and the Fremlin Prize for best thesis.
Further study in Future Energy Systems and Technology through DEWA and UC Berkeley, followed by applied AI and engineering R&D.
My background spans energy research, machine learning, computer vision and edge AI, including work at DEWA’s R&D Centre. I founded h.ai to bring that engineering perspective to practical AI and software.
Today, that means helping construction and engineering teams connect models, quantities, site information and decisions. My ambition is to take the same approach into energy and other industries—wherever intelligent tools can help solve a meaningful problem.
Endurance horse racing, cycling, languages and travel have all shaped how I see the world. I enjoy learning how different people think—and bringing that curiosity back to the things I build.
Start a conversation
Tell us a little about your project. You do not need a technical brief or a particular AI tool in mind.
Three required fields. Company is optional. Please leave out confidential files or sensitive project information at this stage.
Your enquiry goes directly to h.ai so we can respond. It does not subscribe you to marketing.
