How to appoint a PMC in the UAE: an owner's RFP and evaluation framework

Most owners believe the consultancy appointment is decided at the interview. It is decided weeks earlier, in a document nobody outside the procurement team reads closely — and usually by default rather than by choice.

What actually decides a PMC appointment?

The weighting between quality and price, fixed before a single proposal arrives. Everything downstream — the shortlist, the interviews, the scoring, the negotiation — operates inside that decision. Owners who leave it until the envelopes are open have already decided, because by then the numbers are visible and no argument about quality is neutral.

This piece sets out a method an owner can run themselves: how to choose the selection route, how to set and defend the weighting, what the request for proposals should contain, and what the appointment has to lock down afterwards. It is drawn from published institutional practice rather than from opinion, and the weightings are the owner's to set.

Why doesn't UAE law tell you how to weight quality against price?

Because it does not treat consultancy as different from anything else it buys. Federal Decree-Law No. 11 of 2023 requires bids to be evaluated transparently against criteria relevant to the subject matter, and requires the tender announcement to give a clear breakdown of the weight assigned to each criterion. It contains no article specific to consultancy services at all.

Dubai Law No. 12 of 2020 is the same in shape. Article 33 requires financial and technical assessment of bids; Article 37 sets out the grounds for disqualification. Neither prescribes a weighting or a scoring methodology. Abu Dhabi's procurement standards list evaluation criteria and attach no numerical weights to any of them, with services treated alongside goods. There is no UAE mandate for qualifications-based selection of professional services and no UAE standard consultant request for proposals.

The framework does supply one useful signal. Dubai's Administrative Resolution No. 116 of 2025 expressly permits negotiation where works or services require specialised technical expertise or particular professional knowledge, and sets a three-limb prequalification test: technical competence commensurate with the subject matter, adequate financial solvency commensurate with the scope, and good reputation and sound commercial standing. Dubai's own rules recognise that specialised consultancy is not a commodity to be bought on price. They simply do not tell a private owner how to run the competition.

What happens if you let price carry more than twenty per cent?

The competition stops being about quality, whatever the scoring sheet says. FIDIC states the mechanism directly: if the weight allocated to technical qualification is less than eighty per cent, the selection process in most cases automatically turns into cost-based selection.

The convergence on that boundary is unusually tight. FIDIC recommends the financial component be weighted between zero and ten per cent, with twenty as the absolute maximum in exceptionally simple cases, and states that technical qualification should in no case fall below eighty per cent. It is blunt about the alternative: selection on the basis of price alone should never be used for consultancy services. The World Bank sets cost at twenty points out of one hundred as its normal position; the Asian Development Bank uses the same figure. FIDIC's absolute ceiling is the multilateral banks' default.

A mechanic almost nobody explains to owners changes how that ceiling should be read. The financial score is calculated relative to the lowest bid received, not to the owner's budget — the lowest price scores one hundred and every other proposal scores in proportion. So a single reckless low bid does not merely score well itself. It compresses every other bidder's financial score and drags the whole competition toward price. An owner who set price at thirty per cent and then receives one proposal at half the realistic cost has handed that bidder a structural advantage no technical superiority can offset.

The defence must exist before proposals arrive: the owner's own confidential cost estimate. FIDIC treats a proposal abnormally low against that pre-estimate as non-conforming and says it should not be evaluated. Without the estimate, "abnormally low" has no reference point and rejecting a cheap bid is indefensible rather than principled.

Why do technical scores that look decisive often decide nothing?

Because panels cluster. If every shortlisted firm scores between seventy-two and seventy-eight on technical, the technical criterion has no discriminating power left, and price decides the outcome even inside an eighty-twenty process.

FIDIC names this failure precisely: a clustering of technical marks all awarded around the same values neutralises the technical qualification criterion and leads directly to price-based selection by default. The prescribed remedy is a spread of at least twenty points between the highest and lowest score on each criterion. In practice almost nobody applies it, and most owners have never heard of it.

Two supporting disciplines help. Panel members should score independently before the panel meets, so the first strong opinion in the room does not set the anchor. And on a large panel, ranking by averaging each member's rank rather than their raw scores blunts the influence of any single outlier.

The two-envelope sequence is the third control and it is not a formality. Technical and financial proposals arrive in separate sealed envelopes; the technical are opened and evaluated while the financial stay sealed until technical evaluation is complete and announced; proposals failing the minimum technical threshold have their financial envelopes returned unopened. Evaluators who have seen prices cannot un-see them, and technical scores drift to justify a preference already formed. A price opened only after the technical result is locked cannot have influenced it, and the unopened envelope is the evidence.

Private owners in this market are bound by none of this. It is the discipline public institutions impose on themselves, precisely because they know what happens without it.

What should the RFP contain?

Four documents: a letter of invitation, terms of reference, information to bidders, and the full proposed appointment — issued to every bidder at the outset rather than produced after selection. The two items private requests almost always omit are the ones that most improve what comes back.

Publish the evaluation criteria and their weights inside the request itself. This is standard institutional practice and rare in private UAE processes, and it works for a reason: a bidder who knows methodology carries thirty per cent writes about methodology. A bidder who does not, writes about themselves.

Disclose the construction budget. Owners withhold it believing it protects the fee negotiation. What it actually does is guarantee that proposals are resourced against guesses, which makes them incomparable — and the fee gets negotiated anyway, later, with less information on both sides.

The terms of reference should state the expected input of key staff in staff-time, with minimum experience and qualifications, so that proposals can be compared on resourcing rather than on adjectives. And the request should invite comments on the terms of reference. A firm that has read them properly will have some, and that in itself is a data point.

On shortlist size, the institutional position is counter-intuitive and worth following: three to six firms, and fewer for more complex work — three or four on a complex assignment, four or five on a straightforward one. The reason is proposal cost. A serious submission on a complex project is expensive to produce, and if a good firm calculates that its odds of winning do not justify the effort, you receive a thin proposal from exactly the bidder you wanted.

What should the evaluation matrix contain?

Five or six criteria, each carrying a weight the owner sets and publishes, scored independently, with price held at or below twenty per cent and a minimum score spread enforced. The single most consequential design choice sits inside the people criterion.

That choice is how you score personnel. The World Bank weights adequacy for the specific assignment at sixty to eighty per cent within the key-experts criterion, and general qualifications at only ten to twenty. In other words: score the fit to this project, not the length of the curriculum vitae. General experience was already tested when you drew up the shortlist, and scoring it twice quietly doubles its weight while looking rigorous.

The two control rules belong on the face of the matrix, not in a covering note. Price at twenty per cent maximum. A minimum twenty-point spread between highest and lowest on every criterion. Without the second, the first does nothing.

How do you stop the team you were pitched from disappearing?

Write it into the request and the appointment, because it is close to unenforceable afterwards. The established legal test for challenging personnel substitution requires showing that a bidder's proposal was favourably evaluated on the basis of staff it never expected to use — and substitution made reasonably and in good faith is permitted. Challenges brought after award routinely fail even where most of the proposed team never appeared.

The protections are all available at appointment stage. FIDIC's position is that the selected firm should not be allowed to change key staff, or any major item included in its proposal, unless both parties agree that undue delay in selection has made it necessary. The Asian Development Bank permits substitution only where genuinely necessary — ill health, death, or a staff member proving unsuitable — and requires replacements of at least equal qualification, approved by the client.

The mechanism is a key-personnel designation table inside the appointment: named positions, named individuals, continuity for the duration, advance written notice of any proposed substitution, and the owner's approval not to be unreasonably withheld. Define key personnel broadly — any person whose individual action or inaction can affect timely achievement of the performance objectives — and pair the table with committed time allocation per person per month. Names without hours are not a commitment.

How do you check a UAE consultancy before appointing it?

Not through a public register, because there is not one. Dubai Municipality suspended two engineering consultancy offices for six months in 2025, barring them from obtaining permits for new projects, for practices violating approved regulations and ethical standards. The firms were not named, and no public disciplinary register exists at the Municipality, at Trakhees or at the Dubai Development Authority. An owner running a selection today cannot independently establish whether a bidder is under suspension.

It is also worth being clear what registration signifies. Dubai Development Authority registration costs AED 250 a year and requires a valid trade licence — no classification grade, no capability assessment, no insurance requirement. It is an administrative gate, not a credential, and owners who read "registered" as a quality signal are misreading it. Trakhees accreditation is a substantive assessment, but its portal is not publicly searchable either.

So the burden shifts to disclosure, and the request should carry it. Dubai Law No. 14 of 2025 requires the Municipality's registry to hold each engineering office's classification, licensed consultancy scope and registered technical personnel. Any legitimate firm can produce its own entry — ask for it, and ask which of the six licensed categories it holds. Those categories now include engineering advisory offices and engineering audit offices, the latter authorised to conduct third-party review, so an adviser performing technical review in Dubai may be conducting a regulated activity. Whether it is licensed to do so is a fair question at bid stage.

There is also a rating almost no owner asks for. Dubai Municipality has run an Engineering Excellence Initiative since 2017, revised in December 2023 to add an Owner's Opinion standard. It rates engineering consulting offices up to five stars on compliance with laws and building regulations, quality of transaction submissions, dedication to supervising the construction site, absence of technical complaints on the register, and use of building information modelling — with owners feeding the rating through an electronic questionnaire on the Dubai BPS application. The Municipality's stated purpose is to help project owners select the ideal consultant. Ask a bidder for its rating and its complaints record.

One last item, because owners assume the regulator has handled it. No UAE instrument appears to make professional indemnity insurance a precondition to practise. But decennial liability under the Civil Transactions Law is joint and several between engineer and contractor, and any agreement limiting it is void — so professional indemnity cover is the owner's only practical route to recovery. Require the certificate of currency, the limit and basis, confirmation that the retroactive date pre-dates the appointment, cover maintained for a defined period after completion, and notice if the aggregate is eroded by other claims.

What conflicts should the RFP make a bidder disclose?

More than most requests ask. No UAE instrument prohibits a firm from acting in conflicting capacities on the same project. The conflict-of-interest provisions that do exist in UAE procurement law bind the procuring entity's own staff, not the consultant it appoints. The owner closes that gap contractually or it stays open.

Two prohibitions from that policy carry straight into a request for proposals. A firm engaged to provide consulting services for the preparation or implementation of a project, and any of its affiliates, is disqualified from subsequently providing goods, works or services resulting from or directly related to that earlier consulting work — unless the conflict has been identified and resolved acceptably to the client. And consultants hired to prepare terms of reference for an assignment shall not be engaged for that assignment.

Neither is UAE law. Both describe situations that occur routinely on UAE projects, where a firm may write the brief and then bid it, or design the building and then administer its own design under the construction contract.

So the disclosure schedule should ask for five things. Prior involvement by the firm or any affiliate in the project's feasibility, brief, concept, design or specifications. Any intention or eligibility to bid for downstream design, supply or works on the same project. Ownership and affiliate relationships with designers, contractors, suppliers or operators in the chain. Business or family relationships with the owner's own staff involved in the terms of reference, the selection or the supervision. And current appointments whose interests may be adverse to the owner's on this project.

Then the item everyone forgets: the obligation to keep disclosing. FIDIC's duty is not a declaration at submission — it is to disclose at the earliest opportunity and to continuously monitor developments in the firm so that later events do not quietly undermine the stated position. That belongs in the appointment as a standing obligation, because the conflict that matters usually arises after award.

A closing note on the obvious. A piece explaining how to select a consultancy is written by a firm that would like to be selected, and every published attempt at this suffers from it — the criteria tend to describe the author. The only honest answer is a method the owner can run without us, with weightings the owner sets. A matrix that could rank us third is worth more than a list of qualities that happens to describe us.