AI is transforming property, but who can you trust?
Listen · 1-minute summary
0:00 / 1:20
Read the summary
The future of property intelligence is about trust. Artificial intelligence is changing how South Africa's banks and insurers value properties and make lending decisions. At its recent Trust Algorithm conference in Johannesburg, Lightstone asked a key question: can people trust the technology deciding about their homes and finances? Technology analyst Arthur Goldstuck warned that automated actions need to be traceable, explainable and reversible. Lightstone also says approximately 60% of South African households fall outside the Deeds Registry, leaving them largely invisible to traditional property intelligence. Aerial imagery, computer vision and connected databases could help close that gap. AI could also spot financial distress earlier, before a borrower misses a mortgage repayment. Yet human judgement still matters. Sara Winstone, Managing Director of Lightstone Property, said: "We have learned over 20 years of working with banks, insurers and property businesses that trust is built through behaviour over time." The winners in the AI revolution may be those who can show their decisions are accurate, fair, transparent and accountable. Read the full story by Neale Petersen on rei.co.za.

As artificial intelligence begins making decisions about property valuations, home loans and financial risk, Lightstone warns that trust, accountability and human judgement will define the next era of property intelligence.
- AI accountability: Automated property and lending decisions must be transparent, explainable and open to human challenge.
- Hidden property market: Around 60% of South African households remain outside the Deeds Registry, limiting traditional property intelligence.
- Future opportunity: Connected data, predictive analytics and responsible AI could unlock financing opportunities and identify financial distress earlier.
The future of property intelligence is about trust
Artificial intelligence is rapidly changing how South Africa’s banks, insurers, property businesses and financial institutions assess risk, value properties and make lending decisions.
But as AI evolves from analysing information to recommending and executing decisions, a more fundamental question is emerging: Can consumers and businesses trust the technology making decisions about their homes and finances?
This was the central theme of Lightstone’s recent 20 Years in Banking – The Trust Algorithm conference in Johannesburg, where industry leaders examined how data, automation and artificial intelligence could reshape South Africa’s property ecosystem over the next two decades.
For Lightstone, which has spent 20 years providing property intelligence to financial institutions and industry stakeholders, the future is not simply about collecting more data.
It is about transforming information into intelligence that is accurate, transparent, accountable and trusted.

Sara Winstone, Managing Director of Lightstone Property, says trust is built through consistent behaviour and strong relationships.
“We have learned over 20 years of working with banks, insurers and property businesses that trust is built through behaviour over time.”
She says this becomes increasingly important as AI accelerates the speed and scale of financial and property decisions.
1. More data does not automatically mean better decisions
Financial institutions have access to unprecedented volumes of information through sophisticated analytics, automated systems, APIs and AI-powered models.
However, greater data availability does not necessarily translate into better decision-making.
Anthony Miller, Lightstone co-founder and CEO of Simply, highlighted the distinction between accumulating data and producing trustworthy intelligence.
The critical requirements
- Accuracy: Property and financial information must be reliable and regularly validated.
- Transparency: Organisations must understand how automated recommendations are produced.
- Accountability: Decisions should be traceable to responsible systems and people.
- Auditability: Institutions must be able to review and explain decisions.
- Consumer protection: Customers should have appropriate opportunities to question or challenge outcomes.
The message is clear: technology becomes valuable when it improves the quality of decisions, rather than simply increasing their speed.
2. Arthur Goldstuck: AI must remain accountable
Technology analyst Arthur Goldstuck warned that the emergence of agentic AI introduces new challenges for financial and property businesses.
Unlike traditional systems that primarily analyse data or generate recommendations, agentic AI can initiate processes and potentially execute actions with increasing autonomy. That raises important questions about responsibility.
Four questions every organisation must answer
- Who authorised an automated decision?
- Can the reasoning behind that decision be explained?
- Can an incorrect outcome be challenged?
- Can an automated action be reversed?
Goldstuck emphasised that automated actions need to be traceable, explainable and reversible. For the property sector, these principles are particularly important when technology influences mortgage approvals, property valuations, insurance assessments and financial risk.
The greater the authority delegated to AI, the greater the need for accountability.
3. Lightstone data reveals a major property market blind spot
One of the most significant opportunities identified at the conference involves improving visibility across South Africa’s fragmented property ecosystem.
According to Lightstone, approximately 60% of South African households fall outside the Deeds Registry, making them largely invisible to traditional property market intelligence systems. This creates challenges for property valuation, lending and the assessment of household assets.
How technology could close the gap
- Aerial imagery: Identifying properties and physical developments not captured in conventional datasets.
- Computer vision: Analysing buildings, infrastructure and settlement patterns.
- Municipal information: Providing additional insights into properties and services.
- Alternative credit data: Helping financial institutions understand households beyond traditional credit records.
- Connected databases: Combining information from municipalities, banks, valuers and property businesses.
Winstone believes better visibility could create opportunities for households historically excluded from formal property finance.
“Better visibility is not simply a data problem but could help create a path towards formal funding for people whose homes and assets have historically been difficult for the financial system to see.”
For investors and lenders, more comprehensive property intelligence could improve risk assessment while supporting broader financial inclusion.
4. From reacting to financial distress to anticipating it
AI could also change how banks identify and manage mortgage repayment risks.
Currently, financial distress is often recognised only after a borrower misses a repayment. A more connected intelligence system could potentially identify warning signs earlier.
The potential benefits
- Early-warning indicators
Identifying financial pressure before mortgage defaults occur. - Municipal payment patterns
Detecting emerging affordability challenges. - Predictive analytics
Improving understanding of changing borrower circumstances. - Proactive engagement
Allowing lenders to assist customers before problems escalate.
RMB Chief Economist and Head of Research Isaah Mhlanga highlighted the limitations of traditional economic indicators, which can lag developments in rapidly changing markets.
More timely information and AI-enabled analysis could help financial institutions respond sooner, provided the underlying data remains reliable. However, early intervention must be balanced with appropriate privacy safeguards, consumer protection and responsible data use.
5. Human judgement becomes more important, not less
Despite AI’s growing capabilities, industry leaders emphasised that human expertise remains essential.
Josephine Baliah of Nedbank argued that artificial intelligence will not replace human ingenuity, particularly where trust, relationships and judgement influence decisions.
Where people remain indispensable
- Context: Understanding circumstances that automated models may overlook.
- Empathy: Supporting customers facing financial difficulties.
- Judgement: Evaluating complex or exceptional situations.
- Accountability: Taking responsibility when systems produce incorrect outcomes.
- Relationships: Building confidence between institutions, clients and business partners.
AI can process enormous volumes of information, but technology alone cannot replace responsible leadership or human understanding. The strongest systems will combine advanced analytics with meaningful human oversight.
Trust will be the real competitive advantage
The next generation of property intelligence could transform how South Africans buy homes, secure finance, assess property values and manage investment risk.
For Lightstone, the opportunity extends beyond individual valuations and reports towards a connected intelligence ecosystem linking financial institutions, property professionals, municipalities and consumers.
The potential benefits are substantial: more informed lending, earlier identification of financial distress, better property visibility and greater access to formal finance.
Yet these advances will only deliver sustainable value if consumers and institutions trust the systems behind them.
For South Africa’s property industry, the winners in the AI revolution may not be those with the most data or the most sophisticated technology, but those who can demonstrate that their decisions are accurate, fair, transparent and accountable.
The future of property intelligence will ultimately be determined not just by what AI can do, but by whether people trust it to make the right decisions.
Source: Lightstone – 20 Years in Banking: The Trust Algorithm conference, Johannesburg.
Website: www.lightstone.co.za











