Maintenance Contracts and AMC Management: A GCC Focus

Ask a facilities director in Abu Dhabi or Dubai how their maintenance operation is structured and the answer will almost always start with a contract, not a strategy. There is an AMC for the chillers, another for the lifts, one for fire and life safety, one for the BMS, a comprehensive package on the generators, and a labour-only arrangement with a small sub-contractor for plumbing. That stack of AMCs is the real operating model. It defines what work gets done, who pays for a failed part, how fast somebody has to turn up, and what the client is entitled to complain about. And yet in a large share of the organisations I have worked with, the contracts themselves live in a shared folder, the visit schedule lives in a spreadsheet, and nobody can answer the one question that matters most at renewal: did we actually deliver what we promised, and what did it cost us to do it?

The message up front: an AMC is a set of obligations with a fixed price attached. The obligations are only manageable if they are decomposed into an asset list, a visit calendar and a response clock, and only defensible if every visit leaves an auditable record. Where AMC portfolios lose money, the cause is almost never the price in the contract. It is a mismatch between the scope the contract describes and the asset register the team actually maintains.

1. What an AMC actually commits you to

An annual maintenance contract is a fixed-term agreement, usually twelve months, under which a service provider maintains a defined set of assets at a defined location for a defined fee, normally billed monthly or quarterly in advance. That is the whole idea. Everything difficult about AMC management comes from the words "defined" appearing three times in that sentence and being treated loosely in practice.

A workable AMC needs to answer, in writing, a specific set of questions. If any of them are absent, the gap will be filled later by whoever argues more persuasively, which is a poor way to run a commercial relationship:

• Which assets are covered? Not "the HVAC system" but a schedule of assets with tag numbers, make, model, capacity and location. Anything not on the schedule is out of scope by default.
• What work is included? Planned preventive visits at stated frequencies, breakdown attendance, statutory inspections, minor rectification, and the boundary where a job becomes a separately quoted repair.
• Are parts included? This single question separates comprehensive from non-comprehensive and drives most of the price difference.
• What are the response and rectification times? Distinguish the clock to attend from the clock to restore, and set them by priority class rather than one blanket figure.
• What is explicitly excluded? Consumables, refrigerant gas, major component replacement above a value threshold, damage from misuse or power events, civil works, vandalism, upgrades and obsolescence.
• What are the working hours? Whether visits and callouts inside the fee are limited to normal hours, and what an out-of-hours attendance is charged at.
• How is the contract renewed or terminated? Notice periods, price-review mechanism, and what happens to open rectification items at expiry.

What is striking is how often one or two of these are missing from a signed contract, and how reliably the missing one becomes the dispute nine months later. The practical habit I recommend is a one-page contract summary sheet, generated from the contract record, that a supervisor can read in ninety seconds. If the supervisor cannot tell from that sheet whether a gas top-up is chargeable, neither can the technician standing in front of the unit.

2. Comprehensive versus non-comprehensive: the comparison that drives price

The commercial spine of GCC maintenance contracting is the distinction between a comprehensive AMC, where the provider carries the parts risk, and a non-comprehensive AMC, where the client does. Between the two sit labour-only and inspection-only arrangements. The table below sets out how the same obligations land differently under each model.

The pattern that works when pricing a comprehensive AMC is to insist on an asset condition survey before committing. A provider who signs comprehensive cover on old plant they have never inspected has bought an unpriced liability. Clients should expect a genuine premium on older plant, and be wary of a quotation that carries none.

3. Mapping contract scope to the asset register

This is the section I would ask any operations manager to act on first, because it is where AMC portfolios silently leak margin. The contract describes scope in commercial language; the team executes against an asset register in technical language. If those two representations of the same estate disagree, every downstream number is wrong. The failure modes are consistent:

• Assets in the register that are not in the contract. The team maintains them anyway because they are on site and somebody asked. This is unpriced work, and at scale it is the single largest quiet cost in an AMC operation.
• Assets in the contract that do not exist in the register. Usually decommissioned, replaced or never installed. You are being paid for them, which feels fine until the client audits and asks for three years of visit records you cannot produce.
• Quantity mismatches. The contract says forty split units, the register holds sixty-three. One of the two numbers came from a tender document nobody re-verified after handover.
• Classification mismatches. The contract prices "AHUs" as one class; the register distinguishes AHUs, FCUs and FAHUs with very different task content and visit durations.
• Location mismatches. Assets recorded against a building but not a level, room or system, so a technician on a multi-tower site cannot find them and the visit takes twice as long as priced.

The remedy is procedural rather than technical: reconcile the contract asset schedule against the register at mobilisation, and again at every renewal, and treat the reconciliation as a deliverable with a named owner. Where the register itself is weak, fix the structure before you try to attach contracts to it. The parent-child structure, the location hierarchy and the naming discipline are prerequisites, not refinements, and I have written about that groundwork in the asset hierarchy design guide. If you are still deciding what system holds the register at all, the CMMS buyer's introduction is the right starting point.

4. From contract scope to a PPM visit calendar

Once the asset schedule is agreed, the contract has to become a calendar. An AMC promising quarterly servicing on ninety assets across four buildings is not a commitment to ninety visits; it is a commitment to a pattern of attendance at named frequencies, and the skill is turning that into a schedule your team can resource. Keeping strictly to the contract-administration side of it:

• Derive frequencies from the contract, not from habit. If the contract says monthly, monthly is the obligation regardless of what the technical optimum might be. Where the contract frequency and good practice differ, raise it as a variation, do not quietly substitute your own.
• Group assets into visit routes. Clients care about visits, not work orders. A single quarterly visit that covers thirty assets in one tower is one attendance for the client and thirty completions for you. Design the route so both views reconcile.
• Publish an annual planner at mobilisation. A twelve-month grid showing which asset group is visited in which week, issued to the client at the start of the term, converts most scheduling arguments into a diary conversation.
• Allow float, and state it. A visit due in week nine that lands in week ten is normal. Agree a tolerance window in the contract, for example within the calendar month, so compliance is measured against a defensible rule rather than an exact date.
• Separate statutory from discretionary visits. Statutory inspections carry legal and insurance consequences and should never be netted off against routine servicing in a compliance report.

The technical content of each visit, what the task list should contain and how intervals are set, is a different discipline and I will not restate it here. The scheduling mechanics are covered in the guide on how to build a preventive maintenance schedule, and how compliance against that schedule should be measured in PM KPIs and schedule compliance. What belongs in the contract file is the frequency, the tolerance and the evidence rule.

5. Proving the visit happened

In AMC work, the visit you cannot evidence did not happen. That is not a legal claim, it is a commercial reality: at the point of a client audit, a withheld payment or a renewal negotiation, the record is the only version of events that carries weight. GCC clients, particularly government entities, master developers and large facility owners, increasingly audit visit compliance systematically rather than anecdotally.

What a defensible visit record looks like, in ascending order of strength:

• A completed task checklist against the specific asset tag, with readings captured rather than ticked. A checklist of ticks with no values on it invites the assumption that nobody opened the panel.
• Technician identity and timestamps for arrival and departure, ideally captured on a mobile device at the asset rather than transcribed later in the office.
• Client acknowledgement, a signature or an electronic sign-off from a named site representative. This is the single most valuable field in the record and the one most often skipped when the site contact is unavailable.
• Photographic evidence of the asset condition and of any defect raised, attached to the job record rather than sitting in a technician's phone gallery.
• Observations and recommendations, including anything noted as out of scope. A documented recommendation the client declined is a strong position later; an undocumented one is not.
• A monthly compliance pack reconciling visits due, visits completed, visits deferred with reason, and defects raised, issued on a fixed date every month.

The discipline that makes this work is that the job sheet closes at the asset, not at the end of the week in the office. Retrospective completion turns a record into a reconstruction, and experienced auditors spot it immediately from the timestamp pattern.

6. Response and rectification clocks, and how they are administered

Most AMCs carry two clocks that are frequently conflated. The response clock measures from the moment a fault is reported to the moment a competent technician is on site. The rectification clock measures from report to restoration of service. They behave differently, they fail for different reasons, and they should be penalised differently.

The administration questions that decide whether an SLA works in practice:

• What starts the clock? A call to a hotline, an email, a ticket raised in the client's own system, or an entry in yours. Where the client logs faults in their platform and you work in yours, the integration between the two is the SLA, whatever the contract says.
• What pauses it? Waiting for client approval on a quotation, waiting for access to a tenanted area, waiting for a permit, waiting for a part the client agreed to supply. Pause reasons must be defined, or every breach becomes an argument.
• Who classifies priority? If the client sets priority unilaterally, expect priority inflation. If the provider does, expect the opposite. A published classification matrix tied to asset criticality and service impact removes the incentive on both sides.
• What counts as rectified? Full restoration, or a documented temporary measure with a scheduled permanent fix. Both are legitimate; only one of them should stop the clock, and the contract should say which.
• How is a breach evidenced? Automatically from timestamps in the system of record, or manually from a monthly report. Manual breach counting reliably favours whoever compiles the report.

The design of the priority matrix itself, how response and resolution targets are tiered by criticality and impact, is a topic in its own right and I have set out a framework for it in the SLA matrix design guide. The measurement side, which service metrics to report and how to avoid gaming them, sits in the FM KPI framework. This article stays on the contract-administration question: can you prove the clock times from your own records without manual assembly.

7. Penalties, deductions and how disputes actually get settled

GCC maintenance contracts commonly include a deduction mechanism, sometimes framed as liquidated damages, sometimes as a performance-linked withholding from the monthly invoice, sometimes as a points-based scorecard with financial consequences at thresholds. Whether and how any such clause is enforceable depends entirely on the governing law of the contract, the jurisdiction, the free-zone or onshore status of the parties and the drafting itself. That is a question for counsel, not for an operations article, and nothing here is legal advice.

What I can describe is how disputes tend to be resolved in practice, because the pattern is consistent:

• They are settled on records, not on principle. The party with timestamped, signed, complete job history sets the terms of the discussion. The party reconstructing events from memory and spreadsheets concedes.
• They are settled commercially, not formally. Most end in a negotiated adjustment, a goodwill visit programme, or a credit against the next term, because both sides prefer continuity to escalation.
• They are settled faster when raised early. A missed visit flagged in the month it occurred, with a reason and a recovery date, is an operational note. The same missed visit surfacing in an annual audit is a credibility problem.
• They are worst where scope was ambiguous. Almost every serious AMC dispute I have seen traces back to the scope and exclusions section, not to the penalty section.

The practical control is a monthly performance meeting with a fixed agenda: visits due versus completed, open defects, SLA exceptions with pause reasons, chargeable works pending approval, and variation requests. Sign the minutes. A contract administered in monthly increments rarely produces a large dispute.

8. Renewal management and the data you need to reprice

Renewal is where AMC management either pays for itself or exposes the fact that nobody was measuring. The uncomfortable truth is that most renewals in this market are priced by applying an escalation percentage to last year's figure, because the data to price them properly was never assembled. That is a transfer of value from whoever has worse records to whoever has better ones.

To reprice an AMC on evidence rather than instinct, you need, per contract, for the expiring term:

• Actual labour hours consumed, split between planned visits, breakdowns and chargeable works, against the hours assumed in the original price.
• Parts and consumables consumed against the contract, valued, and identified as included or chargeable. On a comprehensive AMC this is the number that determines whether the contract made money.
• Breakdown volume and pattern by asset, which tells you where the plant is deteriorating and which asset groups should move out of comprehensive cover or into a replacement conversation.
• Visit compliance and SLA performance, because a weak record weakens your negotiating position regardless of the underlying cost story.
• Asset register changes during the term, additions, removals, replacements, so the renewed schedule reflects the estate as it now is.
• Out-of-scope work absorbed, quantified. This is the item most providers cannot produce and the one that most often justifies a price movement.

On the renewal calendar itself, the mechanics matter more than they should. Notice periods are commonly thirty to ninety days, and a renewal that is not started until the notice window has opened is a renewal negotiated under time pressure. The pattern that works is a renewal pipeline view with stages, typically data pack assembled, internal price review, client proposal issued, negotiation, signed, with the first stage triggering around one hundred and twenty days before expiry. The broader obligation-tracking discipline behind this, across all contract types rather than AMCs specifically, is covered in contract renewals and obligations.

9. Sub-contractor AMCs and back-to-back terms

Very few FM providers self-deliver everything. Lifts, fire systems, BMS, chillers on proprietary controls, water treatment and specialist equipment are routinely sub-contracted, often to the OEM or its appointed agent. That creates a chain: the client holds you to an AMC, and you hold a sub-contractor to another one. The risk sits in the gap between the two.

The discipline here is back-to-back alignment, and it has to be checked clause by clause rather than assumed:

• Response and rectification times in the sub-contract must be equal to or tighter than the head contract, allowing for your own triage and escalation time in between.
• Scope and exclusions must match. If your head contract is comprehensive and your sub-contract is non-comprehensive on the same assets, you have taken the parts risk without pricing it.
• Term dates should align, or the sub-contract should expire slightly after the head contract, so you are never obligated without cover.
• Evidence standards must be equivalent. If the client requires signed job sheets per visit and your sub-contractor issues a monthly summary, you cannot meet your own reporting obligation.
• Insurance, permits and site access requirements must be passed down explicitly, including permit-to-work compliance where the site operates one.
• Escalation and liability should be proportionate. A penalty exposure on the head contract with no corresponding remedy in the sub-contract is an uncovered position.

Operationally, the point people miss is that sub-contractor visits have to appear in the same visit calendar and compliance report as your own. Sub-contractor performance measurement and scorecards are a separate subject; here, the requirement is simply that their visits and their evidence land in your system of record.

10. Multi-site and multi-client contract structures

Once a provider passes a handful of contracts, structure becomes the constraint. The question is how contracts, sites, assets and schedules relate to each other, and getting it wrong forces years of manual reconciliation.

The structures that recur in GCC practice:

• One contract, one site. The simplest case and the easiest to administer. Common for standalone buildings and single-asset AMCs.
• One contract, many sites, one client. A retail chain, a bank branch network, a school group. Priced as a portfolio, delivered as routes. The administration trap is that compliance has to be reportable per site and in aggregate, and most spreadsheet-based tracking can only do one of the two.
• Many contracts, one site, one client. A large campus or hospital with separate AMCs by discipline. The trap is overlapping and orphaned scope at the boundaries, for example who owns the chilled water pumps feeding equipment covered under a different contract.
• Framework agreement plus call-off contracts. Common with government and master developers: agreed rates and terms at framework level, individual AMCs or work packages called off underneath. Requires the rate card to be held once and inherited, not re-keyed per call-off.
• Shared-service or head-lease arrangements, where an owners' association or master developer holds the contract but individual occupiers raise the faults. Clarify in the contract who is entitled to log a job, and who approves chargeable work.

The data-structure requirement that falls out of all of this is that a contract must be a first-class record in its own right, related to many sites and many assets, rather than an attribute stamped on a work order. Systems that model contracts as a text field on a job cannot answer portfolio questions, and no amount of reporting effort will fix that. The multi-site architecture considerations for FM systems generally are worth reading alongside this if you are operating across several client estates, and how a CMMS handles the facilities case specifically is covered in CMMS for facilities management.

11. What to require from maintenance contract software

AMC management is one of the areas where general-purpose CMMS products vary most widely. Plenty of well-regarded platforms, including several of the mid-market tools such as MaintainX, Limble, Fiix, UpKeep and eMaint, are strong on work order execution and mobile capture while treating the commercial contract as a light reference field. Enterprise asset systems such as IBM Maximo, Hexagon EAM, Infor EAM and SAP PM, and IWMS-class platforms such as Planon, generally have a more developed contract object, sometimes at the cost of configuration effort. The point is not which is better in the abstract; it is that contract management is a capability you must test for explicitly rather than assume.

The requirements I would put in a scoring sheet, and demand to see demonstrated on your own data rather than a vendor sandbox:

On top of the data model, four behaviours are worth insisting on in a demonstration. First, auto-generation of visit work orders from the contract visit plan, forward for at least twelve months, with the tolerance window applied. Second, a visit compliance report that shows due, completed, deferred and reason, filterable by contract and by site, produced without export to a spreadsheet. Third, spares consumption reported against the contract rather than only against the asset, so profitability is visible; the storeroom side of that discipline is covered in spare parts and MRO inventory in a CMMS. Fourth, a renewal pipeline view with stage and owner. If a vendor cannot show all four on a realistic contract during evaluation, assume it will be spreadsheet work after go-live.

12. A mobilisation and administration checklist

This is the short list I would work through in the first thirty days of any new AMC, and revisit at renewal. It is deliberately administrative, because that is where the controllable risk sits:

• Create the contract record in the system of record, with dates, type, value, notice period and billing frequency, before the first visit is scheduled.
• Reconcile the asset schedule against a physical verification of the site, and issue any discrepancy to the client in writing within the mobilisation period.
• Load the visit plan and generate the twelve-month calendar, grouped into routes, with the agreed tolerance window applied.
• Publish the annual planner to the client and obtain acknowledgement of the visit dates.
• Configure priority classes and SLA clocks to match the contract, and confirm what starts and pauses each one.
• Agree the evidence standard for visit completion, including the sign-off route where a site representative may not be available.
• Confirm the exclusion list and value threshold with the operations team in writing, in language a technician can apply on site.
• Check sub-contracts back to back on dates, scope, SLA and evidence standards, and load their assets into the same visit calendar.
• Set the monthly reporting pack contents and issue date, and hold the date from month one.
• Diarise the renewal trigger at roughly one hundred and twenty days before expiry, with a named owner for the data pack.

For reference on wider maintenance and facility management practice, the standards bodies are worth reading directly rather than through vendor summaries: ISO for asset management and facility management frameworks, BSI for the British and European standards commonly referenced in GCC specifications, and ASHRAE for HVAC service and commissioning guidance that frequently underpins MEP scope wording.

The idea to walk away with

An AMC is not a maintenance strategy, it is a commercial container for one, and it has to be administered as a commercial instrument. That means three things held together: a scope that matches the asset register exactly, a visit calendar with evidence attached to every completion, and a consumption record honest enough to tell you at renewal whether the contract made money. Where those three exist, the negotiation is a discussion about facts. Where they do not, it is a discussion about impressions, and the party with worse records pays for the gap.

The AMC is the dominant commercial form in this market. A provider running thirty AMCs is running thirty small businesses, each with its own cost base, obligation set and renewal clock.

Final thoughts

If you take one action from this, make it the asset schedule reconciliation. It is unglamorous, it takes a technically competent person a few days per site, and it is the single intervention that most reliably improves both margin and client relationship, because it removes the ambiguity that every later argument is built on. Everything after it, the calendar, the evidence, the SLA clocks, the renewal pack, is easier once the scope is genuinely known.

On the software question, resist the urge to solve contract administration by buying something. Define the contract data you need to hold, run one contract manually against that definition for a quarter, and only then evaluate platforms against what you have proved you need. You will score vendors very differently after that exercise, and you will recognise a contract module that is a label rather than a capability. AMC terms, remedies and their enforceability vary considerably by jurisdiction and by the specific drafting, so treat everything here as commercial practice rather than legal guidance, and take proper advice on the contract itself.

Related reading: SLA matrix design for FM operations, Asset hierarchy design, How to build a PM schedule, Contract renewals and obligations, Spare parts and MRO inventory in a CMMS, CMMS for facilities management, What is a CMMS.