Nawy’s Chief AI Officer on why AI should show its evidence and not just give answers, why the model is now the easy part, and why the region’s AI should become something it builds on rather than something it showcases.

BB: Real estate is a category where people make some of the biggest, highest stakes decisions of their lives, often with real anxiety attached. How does that shape the way you think about applying AI at Nawy, compared with a typical proptech use case?

Samhaa: The fact that buying a new property is a high stakes decision was actually one of the reasons Nawy was founded. From day one, Nawy has sought to provide its clients with as much transparency and relevant information as possible when making a purchase. That is not particularly easy in a market where information is fragmented, sometimes incomplete or noisy, payment structures are complex, and the consequences of a poor decision may only become visible years later.

Over the years, Nawy has been able to build a significant data layer around the Egyptian market. We receive updated inventory from most major developers on a daily basis, as well as project and developer brochures, floor plans, research and other developer resources. AI has allowed us to process, structure and connect much of this information in ways that would be extremely difficult to do manually.

But having the data is not the same as making it useful to someone making a complex decision. This is where AI becomes particularly powerful. It can interpret a client’s requirements, preferences and constraints, reason across different sources of information, answer complex questions, compare alternatives consistently, and explain why a particular recommendation makes sense.

Our internal Sales Companion AI is an early manifestation of this approach. It supports our property consultants by reasoning across inventory and multiple knowledge bases covering properties, projects, locations and developers. It helps them navigate a market with a breadth and complexity that would be extremely difficult for any individual to hold in their head, and identify the options that are most relevant to each client. Client needs are deduced not only from their verbal requirements. A client who has visited our website or used our app also provides Nawy with signals about what they may be interested in.

In this high stakes category, we also think about AI beyond recommendation. We use AI powered call intelligence to analyse sales conversations, extract client requirements and give managers visibility into interactions that would otherwise be very difficult to monitor systematically. Using AI, we also flag areas where agents may need support, and we are working on extending this to identify potential inaccuracies in the information being communicated to clients.

For us, the value of AI is not in removing the human consultant from the decision. It is in making both the consultant and the client better informed, while also helping us improve the quality, depth and consistency of the advice being given.

BB: What’s one initiative from the past year you’d point to as evidence of where AI at Nawy, and proptech in the region more broadly, is actually heading?

Samhaa: One of the most important directions for us over the past year has been moving from individual AI applications toward what I would describe as a Nawy intelligence layer.

The first generation of AI in many companies, including Nawy, was largely feature driven: build a chatbot, automate a process, score a lead, and so on. Those applications can create real value, but on their own they do not fundamentally change the business. What becomes much more interesting is when AI can reason across the underlying data and knowledge and make that intelligence available across multiple workflows and business lines.

This is particularly relevant at Nawy because Nawy is not a single business operating through a single customer journey. It brings together multiple business lines across the real estate ecosystem, each with its own users, workflows, data and decisions. Nawy’s intelligence layer creates the possibility of building applications that understand not just a property, a client or a transaction in isolation, but the relationships between them, and of making that intelligence available across different parts of Nawy.

That is where I believe proptech should be heading: away from AI as a collection of applications built independently within individual business lines, and toward a shared intelligence fabric connecting data, decisions and transactions across the real estate ecosystem. This can then power much more capable applications and use cases than any individual AI system could in isolation.

BB: Where do you see the biggest gap between how AI in real estate is discussed publicly and how it’s actually being used to bring transparency to buyers and sellers?

Samhaa: Transparency requires accurate data and traceable sourcing. Much of the public discussion around AI in real estate focuses on what sits on top of the data, such as valuation models, recommender systems and conversational engines. Far less attention is paid to the underlying data and knowledge layer that powers them.

In a region where publicly accessible, standardised and sufficiently granular real estate data remains limited and uneven, I believe the biggest gap is often beneath the models: the availability, quality and structure of the data that feeds them, and the ability to connect that data in a meaningful way. Without a trustworthy foundation, however sophisticated the application may appear, its value will ultimately be limited.

But good data alone is not enough. AI needs to make sense of fragmented information by connecting relevant data points, putting them into context, and explaining how they relate to the conclusion or recommendation it produces. The basis of that conclusion should be traceable back to its sources, and where the information is incomplete or the conclusion uncertain, that should be made clear as well.

That is what turns data into transparency. Markets become more transparent when consumers can understand not only what something costs, but what is driving that price, how it compares with relevant alternatives, what trade-offs they are making, and how the economics change under different assumptions.

I believe this is where AI can have a profound impact on real estate in our region: not simply by producing better answers, but by allowing buyers and sellers to understand the information, context and evidence behind those answers.

BB: What’s a lesson from building AI strategy for a trust sensitive, high value category like this that you think other AI leaders underestimate?

Samhaa: One lesson that has become very clear to me is that AI cannot be treated as a separate technology agenda. Once it becomes meaningful, it starts to affect every line of business. Nawy brings together multiple business lines across the real estate ecosystem The role AI can play in our off plan sales business is very different from the role it can play in Nawy Now, our mortgage and Move Now, Pay Later offering; Nawy Shares, our fractional real estate investment platform; Nawy Unlocked, which focuses on finishing, furnishing and monetising underutilised properties; or Nawy Keys, our rental business.

But as AI spreads across the organisation, those efforts cannot evolve independently. They need shared foundations: common data and knowledge, reusable AI capabilities, integration with existing and legacy systems, and a coherent enterprise architecture. AI only becomes transformative when it is integrated into the systems and workflows through which the business actually operates.

At the same time, an AI strategy has to be designed for continuous change. The AI landscape is moving too quickly for us to build an AI system, evaluate it once and consider it finished. The ability to continuously evaluate models and applications, measure how they perform in real business settings, and update or replace components as better capabilities emerge has to be built into the strategy from the beginning. In AI, the ability to evolve the system is almost as important as the ability to build it in the first place.

Just as importantly, it needs shared governance: clear standards around data, model use, evaluation, accountability, human oversight, security, and when AI should or should not be allowed to make or influence a decision. In a trust sensitive, high value category like real estate, governance cannot be something added after deployment. It has to be designed into the way AI systems are built, continuously evaluated and used.

For me, that is the larger lesson: AI strategy is ultimately an enterprise transformation and governance exercise, but it also has to be an adaptive one. The technology matters, but the harder work is ensuring that AI can scale across the business, evolve as the technology changes, and do so without creating fragmented systems, inconsistent standards, or decisions that no one can properly explain or own.

BB: Looking at MENA’s technology and innovation sector as a whole, what’s the shift you expect to matter most over the next two to three years?

Samhaa: I think the shift that will matter most is moving from thinking about AI as an innovation initiative to thinking about it as infrastructure. We have spent the last few years talking about AI applications, pilots and what the latest models can do. I think we are now moving beyond that.

One lesson from the past few years is that advances in AI have increasingly made the model the easy part. Building around that model and on top of it, however, is in no way easy, unless what is being built is trivial. Over the next few years, I expect AI to become increasingly industrialised and to become part of the underlying infrastructure through which businesses and governments operate, much as cloud and digital platforms have become today.

MENA is particularly interesting in this respect because governments across the region have invested heavily in digital and AI infrastructure. But infrastructure is not only compute. Data is equally important. Governments hold enormous amounts of valuable information across real estate, transport, health, infrastructure, demographics, tourism, energy and many other sectors. Making more of that data accessible, structured and machine readable, while putting the right privacy and security safeguards in place, can provide a foundation on which an entire ecosystem of companies and researchers can innovate.

I also think sovereignty will matter, but I would define it more broadly than simply building local foundation models. The more important question is whether the region builds enough capability across data, compute, infrastructure, talent and applications to retain control over how AI is developed and deployed, and to avoid becoming overly dependent on any single model or technology provider.

The real test over the next few years will therefore not be how many AI initiatives we launch, but how deeply AI becomes embedded into the way the region operates and whether we can turn the infrastructure, data and investment being created today into real economic value.

For me, that would represent an important move away from the AI hype: AI becoming less of something we showcase and more of something we simply build on.