Proof in action
Three assets. One governed method.
We ran TerraVest on three assets for a GCC municipality. For each, the owner supplied a plot reference, a location, an intention and a business question. The AI did the rest, and people reviewed every result.
The method
How a run works
1. Opportunity received
A plot reference, a location, an intention and a question.
2. AI understands
Retrieves the plot, its planning rules and its history from your systems of record.
3. AI validates
Checks the envelope, flags gaps, and tags every value with its evidence.
4. AI analyses and compares
Builds scenarios, costs and returns, and tests each for highest and best use.
5. AI recommends
Writes a recommendation from the evidence, with its assumptions visible.
Control point
6. Human decides
An authorised officer approves, amends or rejects. Nothing moves without it.
7. Agentic follow-up
Approved opportunities are published, matched to investors and tracked.
AI supports the decision. People own it.
Plot A · residential
- Plot area
- 56,285 sq ft
- Envelope
- GFA 132,215 sq ft · FAR 2.35 · up to 14 floors
What the owner gave the AI
- Plot reference
- Plot A (anonymised reference)
- Location
- Established residential district, GCC city
- Stated intention
- Bring the plot to market for private development
- Business question
- What is the highest and best use, and is it worth taking to investors?
Nothing else.
What the AI derived
| Field | Value | Evidence |
|---|---|---|
| Plot area | 56,285 sq ft | AI Retrieved |
| Planning envelope | GFA 132,215 sq ft · FAR 2.35 | Actual Source |
| Height limit | Up to 14 floors | Actual Source |
| Envelope check | Plot area × FAR matches permitted GFA | Calculated |
| Field | Value | Evidence |
|---|---|---|
| Scenarios generated | 4 uses tested | AI Scenario |
| Residential sale-price evidence | Comparable transactions, blended rate | AI Estimate |
| Construction cost basis | Benchmark rates plus contingency | AI Estimate |
| Gross development value | AED 595.0M – 661.1M | Calculated |
| Land value basis | To be confirmed by valuer | Requires confirmation |
Scenarios compared
4 scenarios tested
| Scenario | Outcome |
|---|---|
| Villas | Uses a small share of the permitted GFA |
| Mid-density residential | Leaves permitted GFA unused |
| Hotel | Higher operating risk and a longer route to stable income |
| Full-GFA residentialPreferred | 96 residences; strongest return on the permitted envelope |
Returns
| Measure | Value |
|---|---|
| Gross development value (GDV) | AED 595.0 – 661.1M |
| Total development cost (TDC) | AED 453.0 – 457.3M |
| Profit | AED 143.0 – 206.3M |
| Return on cost (ROC) | 32–45% |
| Margin | 24–31% |
| Project IRR | 21.3% |
| NPV at 12% | AED 62.0M |
| Peak funding | AED 327.2M |
The recommendation
Proceed to detailed concept optimisation
What still needs confirmation
- Requires confirmationLand value basis, to be confirmed by an independent valuer
- Requires confirmationSale-price evidence, to be refreshed with current transactions
- Requires confirmationConstruction cost rates, to be validated by a quantity surveyor
- Requires confirmationUnit mix and sizes, to be tested in concept design
We keep this list in on purpose. A recommendation is only as good as the evidence behind it.
Plot B · commercial
- Plot area
- 50,870 sq ft
- Envelope
- GFA 256,000 sq ft · FAR 5.03 · G+29
What the owner gave the AI
- Plot reference
- Plot B (anonymised reference)
- Location
- Commercial corridor, GCC city
- Stated intention
- Offer the plot to private investors
- Business question
- Which commercial scheme gives the best risk-adjusted return?
Nothing else.
What the AI derived
| Field | Value | Evidence |
|---|---|---|
| Plot area | 50,870 sq ft | AI Retrieved |
| Planning envelope | GFA 256,000 sq ft · FAR 5.03 | Actual Source |
| Height limit | G+29 | Actual Source |
| Sellable area | 209,918 sq ft at ~82% efficiency | Calculated |
| Field | Value | Evidence |
|---|---|---|
| Scenarios generated | 3 commercial schemes | AI Scenario |
| Blended sale price | AED 3,700 per sq ft | AI Estimate |
| Construction cost basis | Benchmark rates plus contingency | AI Estimate |
| Gross development value | AED 776.7M | Calculated |
| Strata sales absorption | To be confirmed with market testing | Requires confirmation |
Scenarios compared
3 scenarios tested
| Scenario | Outcome |
|---|---|
| Single-office hold | Capital tied up longer; return depends on leasing |
| Strata sellPreferred | 122 office and retail units, sale-led; highest IRR |
| Hybrid | Partial hold lowers return on capital |
Returns
| Measure | Value |
|---|---|
| Gross development value (GDV) | AED 776.7M |
| Total development cost (TDC)From an earlier run, to be confirmed | AED 555.7M |
| ProfitFrom an earlier run, to be confirmed | AED 221.0M (pre-tax) |
| Return on cost (ROC)From an earlier run, to be confirmed | 39.8% |
| Margin | 27.9%low case, hurdle 15% |
| Project IRR | 24.3%hurdle 12% |
| NPV at 12% | AED 87.3M |
| Peak fundingFrom an earlier run, to be confirmed | AED 310.6M |
| Residual land value vs land value | AED 291.9M vs AED 230.7M |
The recommendation
Proceed to investment review
What still needs confirmation
- Requires confirmationBlended sale price, to be refreshed with current strata transactions
- Requires confirmationSales absorption and phasing, to be tested with the market
- Requires confirmationConstruction cost rates, to be validated by a quantity surveyor
- Requires confirmationTax treatment of profit, to be confirmed with advisers
We keep this list in on purpose. A recommendation is only as good as the evidence behind it.
Plot C · parking
Private paid parking, surface layout
What the owner gave the AI
- Plot reference
- Plot C (anonymised reference)
- Location
- Urban plot, GCC city
- Stated intention
- Operate the plot as private paid parking
- Business question
- Could a private paid parking operation be a viable opportunity here?
Nothing else.
The decision trail
Every step is tagged by where its answer comes from.
1. Understand the plot
Identity, area, boundary, location and current use.
Actual Source2. Read the catchment
Road access, nearby activity, demand signals and competing parking.
Actual SourceAI Estimate3. Design the operation
Capacity, circulation, tariff, hours and grace period.
AI Estimate4. Calculate the case
Revenue, development cost and feasibility screening.
Calculated5. Recommend the next action
Conditional proceed, with a clear validation plan.
Requires confirmation
Indicative feasibility
| Measure | Value | Evidence |
|---|---|---|
| Plot area | 3,729 m² | Actual Source |
| Concept capacity | ≈ 95 bays | AI Estimate |
| AI-selected tariff | AED 10 per hour | AI Estimate |
| Weighted annual occupancy | 35.7% | AI Estimate |
| Indicative annual gross revenue | AED 2.97M | Calculated |
The decision
Conditional proceed
The final decision requires validation.
Validation required before a final decision
- Requires confirmationSite geometry and access
- Requires confirmationMeasured parking demand
- Requires confirmationOperator economics
- Requires confirmationContractor cost
We keep this list in on purpose. A recommendation is only as good as the evidence behind it.
The evidence rule
- Value
- Formula
- Inputs
- Source
- Date
- Confidence
- Status
Worked example: GDV for Plot B
POC run, illustrative figures- Value
- AED 776.7M
- Formula
- Sellable area × blended sale price
- Inputs
- 209,918 sq ft sellable (256,000 sq ft GFA × 82% efficiency) × AED 3,700 per sq ft
- Source
- Sale price: comparable strata transactions. Sellable area: permitted GFA × efficiency.
- Date
- Date of the POC run
- Confidence
- Medium: price evidence is estimated, area is calculated
- Status
- Requires confirmation