How Will AI Transform CRE Investing?

There is no shortage of articles on how AI will change our world and impact virtually every business. It will no doubt have great implications for real estate investment. To examine, I have asked Microsoft’s CoPilot to answer.  (Footnote: Avison Young uses CoPilot as it keeps our searches and data proprietary.) It provides a great overview which I’ve included below only removing the portions on asset management and capital raising. My responses are in red providing my thoughts:

AI will transform commercial real estate (CRE) investing less as a single “disruption” and more as a stack of quiet advantages that compound over time. The firms that adopt it well will underwrite faster, see opportunities earlier, manage assets more precisely, and win capital by appearing smarter and more transparent. Unlike Anthropic CEO Dario Amodei’s prediction that AI could eliminate 50% of entry-level white-collar jobs within one to five years, this response takes a more measured approach. 

Below is a practical, investmentfocused view of where this is heading and why it matters.

1. Underwriting moves from “modeling deals” to “modeling markets”

Today

1. Analysts manually build models. This is very much true where most analysts still rely on static excel models today.
2. Assumptions are based on comps, broker insight, and historical averages. Broker insight remains one of the key differentiators as property specific data is still largely private and relies in many cases on word of mouth.
3. Sensitivity analysis is limited by time and data availability.

With AI

1. AI systems ingest millions of data points: lease-level data, tenant quality signals, local employment trends, mobility data, zoning changes, climate exposure, and even nearby construction permits. Again, this is limited to what data is publicly available. For example, reported cap rates are often projections and subject to projections. Further, lease level data is private.
2. Instead of asking “What’s my IRR if rents grow 3%?”, investors ask: 

“What rent-growth regimes are most likely over the next 36 months, and how does this asset perform in each?” This will be a gamechanger for how investors look to identify markets and asset classes poised for growth. For example, Avison Young is looking at over 50 data sources to identify which multifamily markets are top investment targets moving forward, as opposed to looking at comparable sales in the rear-view mirror. 

Result

1. Faster underwriting (minutes vs. days) This is already happening with analysts leveraging AI assistance to increase their output.  That said, it has not yet replaced more complex modeling. 
2. Better downside visibility
3. Fewer “story deals” and more probabilistic thinking Data driven decision making will become an essential filter for investing. 

Implication for investors: Alpha shifts from spreadsheet skill to question quality and judgment.

2. Deal sourcing becomes predictive instead of reactive

Today

1. Deals come from brokers, relationships, and marketed processes. This is still very much a relationship business where information is not available to all.
2. Off-market sourcing relies on hustle and local knowledge. Off market or “in house” deals that are not publicly listed will be the primary reason that prohibits AI from crossing the moat. Simply put, if brokers continue to keep deals to themselves it will be virtually impossible to disrupt. In my 27 years of brokerage, I have witnessed that many brokers still do not leverage the multiple listings platforms because they want control over the information and who it is sent to. 

With AI

1. Models identify: 
2. Owners under financial stress This can certainly be done comparing debt levels to estimated sales value to uncover underwater properties. That being said, loan maturity dates are one key factor which are not publicly available. 
3. Assets mispriced relative to their micro-market Yes, but without financial information, no two properties are the same. This would be easy to do based on $/SF, but that is not an accurate indicator of value. For example, a property with a long term credit tenant could be worth double what an identical vacant building is worth next door. 
4. Properties likely to trade based on loan maturities, ownership behavior, or operating signals
5. AI flags who to call and why, not just what’s for sale.

Result

1. More true off-market opportunities Yes, if brokers act on this information. I’m still not convinced that owners will engage with an AI chatbot on a multi-million dollar asset. As all brokers know, gaining an owner’s trust can take months, if not years, and many times require an in-person meeting to secure the property information. 
2. Earlier entry into themed trades (e.g., small-bay industrial, medical office conversions)

Implication: The best deals go to firms that combine AI insight + human relationships, not one or the other. I could not agree more. 

3. Valuation becomes continuous and dynamic

Today

1. Values are updated quarterly or annually.
2. Appraisals lag markets. And still rely on publicly available information. I’m not sure if I see a world where appraisal principles change to a more predictive approach as it will introduce subjectivity. 
3. Decision-making often relies on stale snapshots.

With AI

1. Assets have living valuations that update as: 
2. Leasing activity changes. Yes on the revenue side, but on the expense side information is oftentimes only available monthly, quarterly, or even annually.
3. Interest-rate expectations shift. Yes, this would be easy enough to do, but cap rates don’t move lock step with interest rates. 
4. Comparable transactions occur.
5. Portfolio risk can be monitored in near real time.

Result

1. Better sell/hold decisions. Yes, the more real time data that is provided should help.
2. Earlier warning signs on value impairment.
3. More credible communication with capital partners. This is already available with several private platforms like Carta. 

Implication: GPs who offer continuous, data-backed valuation transparency will attract more institutional capital. I agree here. GPs will be able to differentiate themselves by leaning into technology as opposed to making “gut” decisions. 

4. Brokerage and investment sales evolve—but don’t disappear

For someone in investment sales, this matters:

1. AI will: 
2. Instantly benchmark pricing. For the reasons mentioned above, I see this as very difficult to do accurately. 
3. Simulate buyer pools. Yes, AI can identify known investor targets, but it will be difficult to source new capital sources that haven’t made acquisitions yet.  The real power will be with brokers who are able to leverage their investor data across all sale assignments. 
4. Predict execution risk by asset and market.

But it won’t replace: These are all great reasons, why the brokerage industry will be very difficult to replace. 

1. Negotiation.
2. Narrative creation.
3. Capital alignment.
4. Judgment during uncertainty.
5. One other key step to a real estate process is physical inspections. This is another obvious reason for a broker’s involvement. 

The top brokers become:

Interpreters of AI-driven insight, not just distributors of information.

7. The real moat shifts from access to insight

Historically, advantage came from:

1. Access to capital
2. Deal flow
3. Information

AI compresses those advantages.

The new moats

1. Curated proprietary data
2. Clear investment philosophy
3. Ability to ask the right questions
4. Organizational willingness to trust data without worshiping it

8. Risks and limits to be clear-eyed about

AI is not magic:

1. Garbage data → confident wrong answers. This is a key piece to the equation. Those who have access to the best data will prevail in leveraging AI. 
2. Models struggle with regime changes (e.g., COVID, rate shocks).
3. Over-reliance increases herd behavior. I would argue this is already prevalent in CRE investing whether it is data driven or not. 

The winning firms pair AI with:

1. Human skepticism
2. Local market intuition
3. Scenario thinking, not point forecasts

What this means over the next 5 years

1. Short term (1–2 years): Speed and efficiency advantage
2. Mid term (3–5 years): Strategy differentiation and capital attraction
3. Long term: AI-native firms outperform traditional firms structurally

Bottom line

AI won’t replace great CRE investors. It will expose average ones and amplify excellent ones. Excellent conclusion! I could not have said it better myself!