How artificial intelligence could unlock affordable housing

By: Eric Joseph EDUAM

Housing affordability in Ghana and most parts of the world is no longer a theoretical policy concern but a measurable economic constraint with direct implications for household welfare, urban productivity and social stability. International benchmarks commonly define housing as affordable when costs do not exceed 30 percent of household income. Once this threshold is breached, households are forced to trade shelter against essential expenditures such as food, education, healthcare, savings and other essential needs. Although living costs have fallen as of January 2026, low-and middle-income earners in urban Ghana still struggle to meet this index.

Structurally, it is commendable that other macroeconomic indicators, especially the forex rate, have fairly stabilised over the past year. The costs of imported and locally sourced building materials and fuel prices have also seen a downward trend over the same period. Yet, recent estimates put the average price of a supposed “affordable” furnished 2-bedroom residential unit in Ghana at between US$60,000 and US$ 85,000^[1]. This price range remains divergent from prevailing income levels

This article argues that the persistence of high housing costs is not solely a function of macroeconomic pressures, but rather the result of systemic inefficiencies embedded in how housing projects are planned, designed, procured and finally delivered. Within this context, the strategic application of Artificial Intelligence (AI) presents a viable pathway for reducing cost overruns, material waste and coordination failures across Ghana’s housing delivery value chain

Why building costs remain stubbornly high

One principal reason for the high building costs is how Ghanaian buildings are planned, designed procured, and delivered. The issue persists from budgeting through design to project completion. It is particularly challenging to better manage design changes, cost overruns, material waste, and personnel coordination. Small inefficiencies along the value chain compound into significant upward cost shifts at the completion of the building project’s lifecycle. This challenge is prominent when using conventional project management practices.

The conventional project management practices base projections on subjective expertise, fragmented data, and manual computations. These methods are inherently prone to errors and biases. While professional involvement and competence helps to reduce these errors, there is still a systemic risk that the project inputs may be wrongly estimated. This oversight typically incurs higher project costs, longer delays, and at worst, a stall of the building project because the true costs were inadvertently misquoted at the budget phase. At this juncture, I introduce AI as a key means to resolve the problems stressed so that affordable housing schemes can be realised.

Benefits of AI applications in attaining affordable housing

Using AI to design for affordability

Housing affordability is primarily determined at the budget and design stages, where early cost assumptions shape downstream project outcomes. Building owners and project managers must have an honest discourse on the allocated funds for the housing project. This is non-negotiable, as subsequent project phases depend on the outcome of the discussion. After the budget is determined, the design must capitalise on the allocated funds. AI-driven design tools can optimise building specifications, safety compliance and spatial configurations in alignment with predefined budget constraints, thereby reducing over-design and unnecessary material intensity. With proper use of the AI tools, engineers and architects can design projects that meet owners’ specifications and better align material quantities with actual performance requirements.

This step is important in achieving affordable housing objectives. As discussed earlier, marginal changes in concrete volumes, reinforcement steel, and finishing materials can significantly reduce building costs. AI-driven recommendation and predictive modelling systems can tap into local material databases to guide designers in selecting the most cost effective, durable and readily available materials that supports affordability in the long run without compromising quality. This is particularly important in a market heavily reliant on imported building materials.

Predicting and preventing cost overruns

AI-powered applications can also help with more accurate overrun predictions. Where project managers and real estate firms incorporate their historical project information into AI systems, powered by machine learning algorithms, overruns can be reduced to the barest minimum. A key difference between AI systems and traditional manual estimation techniques is that the former can provide sensitivity analysis to demonstrate various conditions under which overruns are likely to occur.

This is a proactive approach to planning for contingencies, which typically contribute to higher unexpected project costs. Predictive capacity is also vital in ensuring that developers and project managers can intervene to adjust specifications, procurement strategies, and construction methods before costs escalate beyond recovery. The insights from the predictive modelling can result in less errors, fewer delays, and more affordable completed housing projects.

Smarter procurement and material management

High building costs in Ghana have a hidden yet strong driving force of procurement inefficiencies. Optimising the selection of suppliers, delivery times, and inventory management can be facilitated by AI. In such settings, materials can be restocked at the lowest possible prices without compromising quality to ensure the project is delivered smoothly. This practice will help reduce direct material, storage and transportation costs.

Computer vision technologies brought to the construction site, leveraging site cameras or drones, are capable of tracking the usage of materials in real time. This technology can detect waste patterns that would not have been noticed otherwise. Even in the context of affordable housing projects with small margins, even slight changes in waste can be significant for final affordability outcomes.

Implications for policy and practice

Although I advocate for AI adoption in the housing and construction sector, I must admit that the affordability crisis cannot be eradicated solely by AI usage. To be clear, I believe that AI presents the opportunity for real estate firms and project managers to streamline budgeting, design, supply chains, and delivery of housing projects. This step is a giant leap in the right direction.

Policy makers also need to rally their support behind AI adoption by establishing protocols for data standards, procurement guidelines, ethical safeguards, and regulatory approval processes  for AI use in the sector, and embedding AI-enabled cost controls within public housing programmes. Without such measures, the dream of affordable housing will remain abstract: a slumber we may never wake up from to face reality.

^[1] UN-Habitat (2021). Ghana Housing Profile

The post How artificial intelligence could unlock affordable housing appeared first on The Business & Financial Times.

Read Full Story