How should an owner manage a hotel Technical Services Agreement in the UAE?

An owner who signed a technical services agreement in 2024 priced brand compliance against a market running at 80.7% occupancy and AED 467 RevPAR. That owner is now delivering the same specification into a UAE market at 57.9% occupancy, with RevPAR down 31.8% year on year to June 2026, roughly 5,000 more Dubai keys landing this year, and a AED 2.5bn government relief package in place. The operator's approval rights have not moved. Every unbudgeted approval condition is now recovered over materially fewer room-nights than the feasibility assumed.

Which makes this a good moment to read the document that grants those rights. The technical services agreement is where the operator's design control lives, and in most hotel developments it is the least-governed document in the stack.

What is a Technical Services Agreement, and what does it actually control?

A technical services agreement is the contract under which a hotel operator reviews and approves the owner's design against its brand standards. It governs who approves drawings, against what standard, within what time, and at whose cost. It is separate from the hotel management agreement, and it is often signed earlier, by different people, with far less scrutiny.

That last sentence is the whole argument. Everything below is the proof.

Owners negotiate the management agreement hard — term, fees, performance test, termination rights. The technical services agreement arrives as a schedule, gets a commercial review of the fee, and is signed. The fee is what owners look at. The approval rights are what cost them money.

Why does it matter that the TSA is signed with a different entity from the management agreement?

Because the entity that approves your design and the entity that runs your hotel are frequently not the same counterparty. Technical services agreements filed in public securities filings show the technical services provider and the management company as separate legal entities within the same group, with separate liability caps and no automatic cross-default between the two agreements.

This is not bad faith. Ring-fencing liability across group entities is ordinary corporate practice. But it has a consequence owners rarely price: a design problem traceable to an approval condition does not automatically become a management-agreement issue. The entity that issued the instruction may be thinly capitalised, and its liability cap may be a fraction of the exposure a single approval condition can create.

Two questions belong at term sheet. What is the technical services entity, and what stands behind it? And what is its liability cap as a proportion of the cost it can instruct?

The drafting response is short: name both entities in one recital, cross-default the two agreements in both directions, and take a parent covenant over the technical services obligations. None of this is exotic. It is absent from most executed agreements because nobody asked.

What does operator approval of the design actually mean — and what does it not mean?

Approval means the design is not inconsistent with the brand. It is not a warranty that the design works. Operators contract to a gross negligence and wilful misconduct standard and expressly disclaim responsibility for the adequacy of what they review. But approval does close the operator's right to reopen — which is where the commercial value sits.

A US federal court decided the point in 2011, in a dispute between an owner and an operator over a hotel built to the operator's review. The court held that the operator's approval of the design constituted a waiver of any right to require changes. It also found that the design guide the operator had repeatedly invoked as the governing compliance standard did not exist as a discrete document, and characterised the operator's conduct in demanding post-approval changes as arbitrary and capricious. Judgment and interest ran to roughly USD 32m.

Read that the right way round. The owner won — after litigation, on exceptional conduct, in a foreign court, fourteen years ago. The lesson is not that owners can sue and win. It is that approval, once given and once final, closes the reopening right. Which puts the entire commercial value of the agreement into two definitions: what constitutes an approval, and when an approval becomes final.

If the agreement contains no design freeze, the operator never has to stop.

Who carries design liability for a UAE hotel once the operator has approved it?

The consultant of record and the contractor. Under Articles 821 to 824 of the Civil Transactions Law, decennial liability for total or partial collapse and for defects threatening the building's stability and safety runs for ten years, binds architect and contractor jointly, and is of public order — it cannot be contracted away. Operator approval does not share it.

Stated plainly, because it is a position most hotel owners in this market have never had put to them: the operator approved the design and carries no design liability for it. The consultant and contractor carry a ten-year statutory liability neither can sign away and the operator's approval does not touch. The owner holds an approval trail with no liability attached to it, and a liability trail with no approval rights in it.

That gap dictates one rule, and it is the most important instruction in this article. Every operator approval condition must be issued to the design team as an owner's instruction, through the owner's own contract — never as a direct operator-to-consultant communication. The moment an operator instructs a consultant directly, the owner has created a seam between the party that gave the instruction and the party that carries the liability. It cannot be closed retrospectively, because the contemporaneous record will show a consultant acting on someone else's direction.

Which puts weight on the record itself: what was instructed, by whom, under which contract, on what date. That is document control, and it is why owner-side approval records belong on a controlled platform rather than in email. In our engagements that record sits in Zepth Core, which handles document control, submittals and site operations from design through handover. Disclosure: PRASOON's founder is also the founder of Zepth; the platform figures cited here are Zepth's own — more than 100 projects delivered and over 50 million square feet under management.

Can the operator change the brand standards after the owner has committed the budget?

Usually yes. Brand standards are typically defined as the requirements the operator “periodically prescribes” — a unilateral variation right sitting quietly inside a definition. The operator can move the standard after the owner has fixed the budget, procured the FF&E and locked the programme.

The limiter worth fighting for is the system-wide implementation test: a brand-standard change binds the owner only once it has been implemented at substantially all comparable hotels in the operator's system, in the relevant region and segment. This is winnable because it is a fairness test the operator's own people can defend internally. If it is genuinely a brand standard, the brand is applying it everywhere. If it is not being applied everywhere, it is a preference — and an owner should not fund a preference.

Four protections belong alongside it. A carve-out for changes required by law, which the owner funds and nobody disputes. An exclusion for changes requiring structural, MEP-capacity or gross-area alteration after a defined design gate. A cost threshold above which a change ceases to be an instruction and becomes a negotiated capital item. And an implementation window keyed to the FF&E renewal cycle rather than to the operator's announcement date.

Then there is deemed approval, the sharpest single redline in the document. In agreements we have reviewed, and in at least one filed publicly, the mechanic runs one way: the owner's silence for a stated period is acceptance, while the operator's silence is nothing at all. Owners should ask why the same mechanic that binds their silence cannot bind the operator's.

There is a fair answer — brand compliance should not be conferred by administrative accident. The workable compromise is a two-tier regime: hard deemed approval on defined low-risk categories such as substitutions within an approved palette and coordination-level revisions, and on everything else an escalation route with agreed programme relief when the operator misses its window.

One practical note. Brand standards manuals run to hundreds of prescriptive pages, are revised periodically, and the owner's team is expected both to have read all of it and to know which version governs. Document-reading agents are genuinely useful here: Zepth AI classifies and retrieves against the current manual version and flags where a submission departs from it. The limit is worth stating honestly — this is classification and retrieval, not prediction. It tells you what the document says and where the conflict is. What to concede remains a human judgement.

What does the regulator require that the brand standard does not?

Minimum room and bathroom areas by star rating. The owner's area programme must satisfy the higher of the classification authority's minimum and the brand's standard — and the technical services process tests compliance with the brand only. A design can pass operator review and fail classification.

The Abu Dhabi classification manual fixes five-star minima at 30 m² for a single room, 32 m² for a double, 48 m² for suites and 4.5 m² for bathrooms. Dubai operates its own regime under the emirate's hotel licensing and classification framework. An owner developing in both emirates should not assume one area programme serves both.

The cost point owners discover too late is that area drives cost before specification does. Midscale hotels run at roughly 60 m² per key; a luxury resort runs to around 210 m² per key — a 3.5x range in gross area before anyone has selected a stone. And roughly 40% of a hotel's gross area is back-of-house: space the owner funds, the operator specifies, and the guest never sees.

So the brand's area schedule is a cost instruction, and it should be priced and signed at term sheet, not discovered at concept design. An owner negotiating the area programme before signature is arguing about the 3.5x. An owner negotiating finishes after signature is arguing about the margin on top of it.

This is also the answer to a question owners reasonably ask: if the operator is reviewing the drawings, why appoint an owner's representative? Because the operator reviews against one of the three standards that apply to the building.

What happens to brand-standard cost after the hotel opens?

It continues. Operators keep prescribing standards for the life of the management agreement, funded from an FF&E reserve the owner contributes to as a percentage of revenue. In a market where revenue has fallen sharply, a reserve struck as a percentage of revenue underfunds a renovation cycle whose cost has not fallen with it.

That is the squeeze, and it is where the operating market actually bites. Reserve contributions track RevPAR down. Brand-mandated renovation scope does not. An owner who set the reserve at a default percentage in a record year is accruing against 2026 revenue toward an improvement plan priced on 2024 assumptions.

The discipline is to forecast the renovation cycle against the projected reserve balance rather than discovering the gap when the improvement plan lands, and to benchmark capex per key across the portfolio so any single property's demand can be tested against comparable assets rather than accepted in isolation. That work — capex requests and approvals, budget variance flagged before period close, property-to-property benchmarking and the reporting above it — is what Zepth Edge does, on the operations side of handover.

What should an owner fix in the TSA before signing?

Eight things, every one cheaper to fix at term sheet than at any point afterwards. The register below is the working document: it runs from term sheet through opening and gives the owner's team one place where every obligation, approval trigger and response window is recorded against the entity that owes it.

And the eight redlines:

1. Cross-default the technical services agreement and the management agreement in both directions; name the technical services entity and covenant it.
2. Define what constitutes an approval, and define when an approval becomes final.
3. Make deemed approval two-tier, and make it run in both directions.
4. Apply a system-wide implementation test to every brand-standard change.
5. Set a design gate after which structural, MEP-capacity and gross-area changes are negotiated capital items.
6. Price and sign the area programme at term sheet, tested against the emirate's classification minima.
7. Route every operator condition as an owner's instruction through the owner's own contracts.
8. Test the FF&E reserve against a forecast renovation cycle rather than accepting a default percentage.