Responsible AI in Real Estate should become a professional standard for Kenyan property agents as artificial intelligence becomes increasingly integrated into marketing, client communication, property analysis and everyday sales operations.
An estate agent can now use AI to draft a property description in seconds.
AI can summarise a conversation with a buyer, prepare a follow-up message, organise CRM notes, compare investment options, analyse spreadsheets, generate social-media ideas and virtually furnish an empty apartment.
These tools can make a real estate business considerably more efficient.
But speed creates responsibility.
What happens if an AI tool invents an amenity that the property does not have?
What happens if a virtually staged photograph makes an apartment appear larger or more luxurious than it really is?
What happens if an agent uploads a client’s ID, title deed, bank statement or sale agreement into an AI platform without knowing how that information will be processed?
And what happens when artificial intelligence starts making decisions about which clients deserve attention?
These questions are no longer theoretical.
Kenya launched its National Artificial Intelligence Strategy 2025–2030 in March 2025. The strategy is built around three pillars: AI digital infrastructure, data and AI governance, and AI research, innovation and commercialisation. It also places emphasis on ethical oversight, cybersecurity and public trust.
Kenya National AI Strategy 2025–2030
At the same time, estate agents operate within an already regulated profession.
The Estate Agents Registration Board says its mandate includes ensuring that the competence and conduct of practising estate agents remain sufficiently high to protect the public.
Artificial intelligence therefore does not remove professional responsibility.
It increases the need for it.
The principle for agents should be simple:
Use AI to improve professional judgement, not to replace professional accountability.
Responsible AI in Real Estate: Why Agents Need a Code of Conduct
The real estate industry has adopted AI faster than many agencies have developed rules governing its use.
An individual salesperson may open an AI tool today and immediately start using it to process:
- Buyer conversations
- Property details
- Client documents
- Financial calculations
- Marketing photographs
- Internal company information
The technology makes this easy.
That does not mean every use is appropriate.
Kenya’s Office of the Data Protection Commissioner has specifically warned that increasing AI adoption can create privacy issues involving excessive collection of personal information, lack of informed consent and unauthorised use of data. The ODPC has emphasised the need for technical and organisational safeguards and, where appropriate, Data Protection Impact Assessments.
Office of the Data Protection Commissioner guidance on AI and data protection
A professional agency therefore needs rules before problems occur.
Responsible AI in Real Estate means deciding where technology can assist, where human review is compulsory and where certain information should not be placed into an AI system at all.
1. Verify Every AI-Generated Property Fact Before Publishing
The first rule is the most important:
AI output should be treated as a draft until a human verifies it.
Generative AI systems can produce confident answers that are incorrect.
An agent may provide basic property information such as:
“Three-bedroom apartment in Westlands, heated swimming pool, 210 sqm, completion June 2027.”
An AI tool may rewrite this beautifully.
But if the original information is incomplete, the system could accidentally introduce additional claims.
Suddenly the listing says:
“Three-bedroom + DSQ, 210 sqm, heated pool, rooftop running track and private balcony.”
Perhaps there is no DSQ.
Perhaps the pool is not heated.
Perhaps there is no rooftop running track.
The wording may sound professional while being factually wrong.
Before publication, agents should verify:
Property size.
Bedroom configuration.
Price.
Completion date.
Amenities.
Parking allocation.
Developer information.
Payment plan.
Location.
Rental projections.
Every material property claim should come from an identifiable source.
That could be:
- Developer documentation
- Owner instructions
- Floor plans
- Site inspection
- Official records
- Signed mandate
The AI should organise information.
It should not create facts.
At Realty Boris, the practical standard should be:
If an agent would be uncomfortable saying the statement directly to a client during a physical viewing, it should not appear in AI-generated marketing.
2. Protect Client Data Before Using AI
One of the greatest risks surrounding Responsible AI in Real Estate is not marketing.
It is data.
Estate agents regularly handle highly personal information.
A buyer may provide:
- National ID
- Passport
- Telephone number
- Email address
- Address
- Employment information
- Financing information
- Bank statements
A property owner may provide:
- Title documents
- Lease documents
- Identification
- Company records
- Rental statements
- Tenancy agreements
Agents should not assume they can copy this information into any AI platform simply because doing so is convenient.
Kenya’s Data Protection Act gives individuals significant rights relating to how personal information is processed. It also contains specific provisions concerning decisions based solely on automated processing where those decisions create legal or similarly significant effects.
Before using client information with an AI service, agencies should understand:
What data is being uploaded?
Why is it necessary?
Where is it stored?
Can the provider retain it?
Can the information be used to improve the provider’s models?
Who can access it?
How long is it retained?
An agent does not need a client’s full passport to ask AI to rewrite a follow-up email.
Data minimisation should be the default.
If a task can be completed without personal identifiers, remove them.
3. Clearly Disclose AI-Generated or Virtually Staged Property Images
AI image tools present enormous opportunities for property marketing.
An empty apartment can be virtually staged.
A buyer can see how a bedroom might look with furniture.
A developer can communicate an interior concept before completion.
These uses can be useful.
The problem begins when visual enhancement changes the property’s actual characteristics.
An agent should not use AI to make a property appear to have:
A bigger balcony.
A swimming pool that does not exist.
A different view.
More windows.
Higher ceilings.
Additional landscaping.
A larger living room.
Better finishes.
A neighbouring building removed from sight.
If the buyer believes an AI-created feature physically exists, marketing has crossed from visualisation toward potential misrepresentation.
A professional code should distinguish between:
Editing for presentation and editing reality.
Brightness correction is different from changing the view.
Removing a dustbin is different from removing an entire neighbouring apartment tower.
Adding furniture for clearly disclosed virtual staging is different from creating a larger room.
The best rule is simple:
If AI materially changes what the buyer sees, disclose the change.
A label such as:
“Virtually staged image”
or
“AI visualisation – furniture not included”
protects both the client and the agency.
4. Human Agents Must Remain Responsible for Property Advice
A buyer may ask:
“Should I buy the KSh 12 million apartment in Westlands or the KSh 9 million apartment in Kilimani?”
AI can help organise the comparison.
It can calculate:
- Price per square metre
- Estimated rental yield
- Payment schedule
- Potential cash flow
But the final advisory conversation requires professional judgement.
An experienced agent may know that:
The Westlands project has stronger rental demand.
The Kilimani development has hundreds of competing units.
The Westlands developer has a stronger completion record.
The Kilimani apartment has a much better floor plan.
These contextual details may never appear in the AI’s dataset.
Responsible AI in Real Estate therefore requires a human professional to remain accountable for recommendations made to clients.
AI can support the advisor.
It should not become the advisor.
5. Do Not Present AI Calculations as Guaranteed Investment Returns
Investment analysis is one area where AI can appear particularly convincing.
Suppose an agent enters:
Purchase price: KSh 10 million.
Expected monthly rent: KSh 100,000.
AI calculates a gross annual rental yield of 12%.
The mathematics may be correct.
The assumption may not be.
Can the apartment actually achieve KSh 100,000?
Will it remain occupied for 12 months?
What is the service charge?
Will management fees apply?
Does it need furnishing?
What about repairs?
A sophisticated analysis should separate:
Known figures
from
Assumptions
from
Forecasts.
Agents should never allow an AI-generated projection to become:
“Guaranteed 12% return.”
That language can give clients a false impression of certainty.
The better approach is:
“Based on an estimated monthly rent of KSh X, the projected gross yield would be approximately Y%, before expenses and vacancy.”
The agent should then explain the assumptions.
AI can improve calculations.
It cannot eliminate investment risk.
6. Avoid Unfair Automated Client Profiling
Real estate CRMs increasingly use automation to score leads.
This can be useful.
An agency with 500 enquiries needs to identify:
- Ready buyers
- Long-term prospects
- Investors
- Tenants
- Sellers
But automated scoring can create problems if the criteria become inappropriate.
For example, an AI system should not automatically assume that someone deserves lower-quality service because of personal attributes unrelated to their actual property requirements.
Kenya’s Data Protection Act addresses automated individual decision-making where decisions based solely on automated processing significantly affect a person. It provides safeguards including notification and the ability, in certain circumstances, to request reconsideration.
For estate agencies, the practical rule should be:
AI may help prioritise workflow.
It should not silently determine a person’s worth as a client.
Useful criteria might include:
- Budget
- Purchase timeline
- Requested location
- Viewing activity
- Response history
But agencies should be cautious about systems that make assumptions using sensitive or irrelevant characteristics.
7. Responsible AI in Real Estate Must Include Fraud Prevention
The emergence of AI also changes how property fraud may look.
Fraudulent communication no longer needs to contain poor grammar, suspicious formatting or obviously fake imagery.
AI can create highly professional:
- Emails
- Voice messages
- Documents
- Images
- Videos
Kenya’s National KE-CIRT/CC has reported that social-engineering attacks increasingly involve AI-generated deepfake scams and voice impersonation. Its cybersecurity reporting notes that AI-generated voices have been used to impersonate executives and facilitate fraudulent transfers.
Communications Authority cybersecurity reports
Imagine a property buyer receives a WhatsApp voice note that sounds exactly like their agent:
“We have changed the bank account. Please send the balance to this new account today.”
Five years ago, recognising someone’s voice might have felt like strong verification.
That is becoming weaker evidence.
Professional agencies need stronger payment-verification procedures.
For example:
A change of bank details should never be communicated solely through WhatsApp.
Large payments should be confirmed through established official channels.
Clients should know the company’s authorised accounts.
Sensitive transaction changes should require secondary verification.
As AI-generated impersonation becomes more convincing, trust needs to depend increasingly on process rather than appearance.
8. Keep Important AI-Assisted Communications on Record
AI can generate dozens of client messages quickly.
That convenience can create another problem:
Nobody remembers what was promised.
If an automated assistant tells one buyer:
“The project will definitely complete in June.”
and another:
“Completion is September.”
the business has created inconsistent representations.
Important communication should therefore remain recorded.
Agencies should retain appropriate records of:
- Property specifications
- Price information
- Offers
- Payment instructions
- Material client representations
- Important automated responses
The goal is not to record every casual conversation forever.
The goal is accountability.
When AI is involved, the company should still be able to answer:
What information did we give the client?
Where did that information come from?
Who approved it?
This becomes particularly important for high-value transactions.
9. Train Agents Before Giving Them AI Tools
Many AI problems are actually training problems.
An estate agent may understand sales extremely well but know little about:
- AI hallucinations
- Prompt privacy
- Deepfakes
- Data retention
- Automated profiling
- Cybersecurity
Giving employees access to powerful AI tools without guidance can expose both clients and the agency.
Kenya’s National AI Strategy explicitly includes AI skills, governance and responsible deployment within the country’s broader AI ambitions.
A real estate agency should therefore train employees on practical questions:
Which AI tools are approved?
What client data may be entered?
What information is prohibited?
Who checks AI-generated marketing?
How should virtual staging be labelled?
Can AI calculate returns?
Who verifies those calculations?
What happens if an AI response is wrong?
Training turns AI from an informal experiment into a professional business tool.
10. Create an Agency-Wide AI Policy
The final rule of Responsible AI in Real Estate is governance.
AI should not operate according to whichever rules individual agents invent for themselves.
A property company should establish one internal standard.
This is particularly important for larger teams.
Without a policy:
One agent may disclose virtual staging.
Another may not.
One may upload client IDs.
Another may avoid it.
One may review every AI-generated description.
Another may publish immediately.
That inconsistency creates risk.
An internal AI policy should define:
Approved uses: marketing drafts, summaries, research support, CRM assistance and administrative work.
Restricted uses: sensitive documents, client financial data, identity documents and confidential contracts unless approved safeguards exist.
Required human review: property descriptions, investment calculations, market claims and client advice.
Disclosure requirements: virtual staging, materially AI-generated visuals or simulations.
Security procedures: payment changes, account verification and suspicious communications.
Accountability: every AI-assisted output still has an identifiable human owner.
The policy does not need to be 50 pages.
It needs to be clear enough that every employee understands the boundaries.
Responsible AI in Real Estate and Professional Accountability
Estate agents should remember that AI has not replaced Kenya’s professional regulatory framework.
The Estate Agents Registration Board regulates estate-agency practice in Kenya and states that only registered and licensed estate agents holding current practising certificates may conduct estate-agency practice.
Estate Agents Registration Board
EARB’s strategic framework says its role includes ensuring professional competence and conduct sufficiently high to protect the public.
The Estate Agents Act also provides for a professional code of conduct and regulation of practice.
That matters because technology does not create a separate professional identity.
The licensed professional remains responsible.
An agent should never respond to a serious error by saying:
“The AI wrote it.”
If you published it, sent it or relied on it professionally, you remain responsible for checking it.
Where Estate Agents Can Use AI Productively
A strong code of conduct should not discourage useful technology.
Agents can benefit substantially from AI when it is applied correctly.
Good uses can include:
Property descriptions: drafting copy from verified property information.
Follow-up messages: creating personalised drafts after a viewing.
CRM summaries: turning long conversation histories into clear notes.
Meeting notes: summarising client requirements.
Market comparisons: organising verified property data into comparable tables.
Content ideas: creating educational topics for blogs and social media.
Frequently asked questions: drafting initial responses for common enquiries.
Translation: supporting communication with international clients, followed by human review where accuracy is important.
Spreadsheet analysis: identifying patterns in leads, viewings, offers and conversions.
The common factor is that AI assists the professional rather than pretending to replace them.
AI Practices Estate Agents Should Avoid
A clear code also requires red lines.
Agents should avoid:
Fabricating property information.
Publishing AI-generated descriptions without checking them.
Creating fake client testimonials.
Using AI to impersonate another professional.
Altering property imagery deceptively.
Uploading confidential client data unnecessarily.
Allowing automated systems to make important client decisions without appropriate safeguards.
Creating fake documents.
Presenting hypothetical returns as guarantees.
Using AI to conceal property defects.
Automating payment changes without human verification.
These practices undermine trust whether or not legislation specifically mentions artificial intelligence.
A Practical Realty Boris AI Code of Conduct
At Realty Boris, a practical approach to Responsible AI in Real Estate could be summarised through ten commitments:
- Accuracy first: every material property fact must be verified.
- Client privacy: personal information is never uploaded unnecessarily.
- Visual honesty: AI staging and material image alterations are disclosed.
- Human accountability: agents remain responsible for advice.
- Transparent projections: investment forecasts clearly state assumptions.
- Fair treatment: AI does not determine client worth using inappropriate profiling.
- Fraud prevention: payments and identity-sensitive instructions require verification.
- Record keeping: important representations are documented.
- Staff training: team members understand both AI capabilities and risks.
- Professional standards: technology never overrides legal, regulatory or ethical duties.
This creates a simple operating principle:
AI may increase speed. It must not reduce trust.
Frequently Asked Questions
Can Estate Agents in Kenya Use Artificial Intelligence?
Yes. AI can assist with marketing, CRM management, communication, analysis and administration.
However, existing obligations around professional conduct, client protection and data privacy still apply.
Should Agents Tell Buyers When an Image Is AI-Generated?
If AI has materially altered or virtually staged the property, disclosure is the more transparent professional approach.
The client should understand what physically exists and what is a visualisation.
Can I Upload a Client’s ID or Passport to an AI Tool?
Agents should be extremely cautious.
The ODPC has stressed that data-protection obligations continue to apply when AI systems process personal information, including requirements around lawful processing, safeguards and data-subject rights.
Can AI Give Property Investment Advice?
AI can support calculations and comparisons, but agents should verify inputs and provide appropriate human judgement.
A projected return should never be presented as guaranteed merely because an AI system calculated it.
Can AI Replace an Estate Agent?
AI can automate many administrative tasks, but estate agency also requires negotiation, local knowledge, client relationships, property inspection and professional accountability.
EARB continues to regulate authorised estate-agency practice in Kenya.
What Is the Biggest AI Risk for Property Agents?
There is no single risk.
The most important areas include inaccurate information, privacy breaches, misleading images, poor automated decisions and AI-enabled fraud.
Realty Boris Expert View on Responsible AI in Real Estate
At Realty Boris, we believe Responsible AI in Real Estate will eventually separate strong professional agencies from businesses that simply adopt every new technology without considering the consequences.
AI should help an agent understand the client better.
It should help them respond faster.
It should improve reporting.
It should help organise information.
It should make repetitive work more efficient.
But it should never become an excuse for lowering professional standards.
Real estate is different from many low-value online transactions.
A client may be trusting an agent with:
KSh 10 million.
KSh 30 million.
KSh 100 million.
Or a family home that represents most of their lifetime wealth.
That requires accountability.
The future property agent should therefore combine:
technology + market knowledge + verification + ethics + human judgement.
The most valuable professional will not necessarily be the agent using the most AI.
It will be the agent who knows exactly where AI adds value and where human responsibility must remain in control.
Conclusion
Responsible AI in Real Estate is ultimately about protecting trust while taking advantage of better technology.
Artificial intelligence can make Kenyan estate agencies faster, more analytical and more responsive.
It can help with property marketing, lead management, client communication, investment calculations, reporting and administration.
Kenya’s National AI Strategy shows that AI adoption is becoming part of the country’s wider digital-development agenda, while regulators such as ODPC continue emphasising privacy and responsible data processing.
At the same time, the Communications Authority has highlighted increasingly sophisticated AI-enabled threats such as deepfake scams and voice impersonation.
This means real estate agencies cannot think only about what AI can do.
They must also ask:
Should we use it for this?
What information are we giving it?
Has the output been verified?
Could it mislead a client?
Who remains accountable if it is wrong?
The answer to that final question should always be clear.
The professional remains responsible.
AI should help agents deliver better service.
It should help buyers receive clearer information.
It should help businesses become more efficient.
But accuracy, transparency, privacy and professional integrity must remain non-negotiable.
That is what Responsible AI in Real Estate should mean for the future of Kenyan property agency.
Call to Action
Looking for professional, transparent property guidance in Nairobi?
Contact Realty Boris for expert property advice and carefully selected investment opportunities.
