Asset
Two-floor Class A office~2,300 m² in a new mixed-use tower
A single income-producing commercial office floorplate offered off-plan by a top-tier regional developer in a major GCC financial free-zone. The seller's broker supplied a return model showing a high-teens IRR. The fund rebuilt it from scratch.
The seller's narrative is coherent and the asset is genuinely well-positioned in a supply-starved submarket. But on the fund's own assumptions the return is structurally dependent on two unproven inputs: a new-build rental premium with no signed-lease evidence behind it, and a 30%+ capital appreciation at exit that requires cap-rate compression to the very bottom of the seller's own stated range. Roughly four-fifths of the modelled profit is capital gain, not yield. We do not hide this — we flag it and gate the deal on evidence.
The same engagement any connected user can run on their own deal. Each step is an independent specialist; the final step is an adversary who never saw the work that produced the recommendation.
Illustrative, de-identified figures — proportional to the real deal, shifted so it cannot be traced.
Four unlevered scenarios built from the fund's own assumptions — not the seller's. Each carries a full income-and-exit cash flow including a leasing void and the marketing/concept spend the seller's own analysis admits is required but left out of its model. The base case is highlighted; it is the case the deal must stand on before any leverage is layered.
| Scenario | Rent (rel.) | Exit (rel.) | Void | 2-yr ROI | Unlevered IRR | vs 15% hurdle |
|---|---|---|---|---|---|---|
| Bullseller's inputs | 100 | 100 | 6 mo | ~43% | ~20% | +500 bps ✓ |
| Basefund central case | 90 | 86 | 9 mo | ~21% | ~10% | −500 bps ✗ |
| Bear | 79 | 74 | 15 mo | ~−1% | ~0% | capital loss ✗ |
| Severe bear | 71 | 67 | 18 mo | ~−11% | ~−6% | capital loss ✗ |
The seller's model is internally consistent at the headline — but the rebuild surfaced three things that move the decision.
Re-deriving the cash flows from the seller's own structure produced an unlevered IRR about ~140 bps lower than the headline they quoted. The most likely cause is an undisclosed assumption about partial leasing in year one. The number is not wrong by accident — it is optimistic at every fork. We carry the corrected figure forward, not the brochure's.
The headline exit value is only reachable at the bottom of the "conservative 8–10%" cap-rate range the seller themselves quote — i.e. the single most optimistic point in their stated band. Push the cap rate to the middle of that same range and the exit value drops materially; push it to the top and the asset exits below entry. A "conservative" range whose conclusion only works at its most aggressive end is not conservative.
The seller's SWOT explicitly states a tenant concept must be created and marketing must be funded to lease the building — then leaves both costs out of the return model. The model also counts only one year of rent across a two-year hold (an implied leasing void) without disclosing it. We added the marketing/concept spend and a realistic void to every scenario. The one verification that passed cleanly: the entry net yield of ~9.5% reconciles — the asset is a sound yield play; it is the appreciation thesis on top that is unproven.
Where the independent challenge from Risk pushed back hardest.
The submarket is supply-starved today — which is the whole bull thesis — but a large new Class A scheme is scheduled to deliver in the same year the fund would be trying to sell. New competing supply at exit suppresses cap-rate compression and hands buyers optionality, both of which cut the exit price precisely when the deal is most exposed. The seller frames a nearby prestige development purely as demand-accretive; the fund's read is that newer, taller, higher-grade stock also competes for the same premium tenants the deal needs at its target rent.
Roughly four-fifths of the modelled return is capital gain, not rent. If the appreciation thesis fails, the yield alone returns about its entry yield over the hold — comfortably below the 15% mandate. The bull case also requires three optimistic inputs to hold simultaneously (high rent, high exit, short void); they are correlated to the downside, so the realistic distribution is skewed worse than a naive midpoint suggests.
As a standalone position this asset would represent the fund's entire deployment until further deals are added — a concentration flag in its own right. And there is no identified tenant pipeline for a full-floor block of this size; first-tenant leasing in a brand-new tower typically demands rent-free periods and fit-out contributions that further erode year-one cash flow.
Every open item, graded by how much it moves the decision. The two critical gaps are stop-the-line: the deal cannot proceed to commitment until they are closed with primary evidence.
The single largest return driver. No closed-transaction registry evidence supports the assumed exit value. At a mid-range exit cap rate the value falls sharply; the bull number needs the most optimistic cap rate in the seller's own band. Requires land-registry comparables before any term sheet.
The target rent sits in the upper percentile of transactions for stock that is 12–17 years old. There is no signed-lease comparable from a recent Class A building of this grade — the premium is an assumption, not evidence. Requires registry lease data.
Admitted in the seller's own SWOT, excluded from the return. Must be quantified — its omission inflates the stated ROI.
LTV, rate and tenor unconfirmed, so the levered IRR is uncomputable. At least two indicative term sheets required before the financial model can be finalised.
The handover date differs between the offer table and the project body text. Developer written confirmation required before engagement.
An unexplained amount in the deposit tranche needs a documented cost breakdown; the seller-commissioned valuer's professional accreditation and any conflicts to be confirmed at KYC.
The asset is real and well-located and the entry yield is sound, so the opportunity is not dismissed. But on the fund's own central case it misses the mandate by roughly 500 bps, and the gap to the seller's case rests entirely on two unproven inputs. The fund proceeds to evidence-gathering, and capital is gated behind these conditions — every one of which must clear:
You bring the deal and the AI; the fund brings the discipline. The same eval.full engagement that produced this memo — independent valuation, model forensics, and a mandatory red-team challenge — runs on your inputs, in your jurisdiction, with your sensitive figures never leaving your machine. We return the method; your AI does the synthesis.
Disclaimer. This is an anonymized illustrative sample produced for demonstration. All identifying details have been removed and all figures rounded, indexed or shifted; nothing here is a transaction record, valuation opinion, or investment advice, and no real counterparty, asset or location is identifiable from it. Methodology shown is the platform's analytical process, not a recommendation to buy or sell any security or asset.