Following our article on specification, or lack of it, at Ian McGlinn's Maison d'Or 'Yes! We have no specifications', John Gelder looks at what the case tells us about inspection.

The court made the point that the role of the architect in inspecting works 'has been the subject of surprisingly few cases', so the findings in this case will be a very useful addition to the body of knowledge.

Inspection without specification

It was agreed that design and inspection, together, had the aim of producing work, in this case, 'of a very high standard'. So failure to supply a specification (a feature of this project) was not fatal in itself, but had the consequence that 'periodic inspections to be made by [the architect] were of particular significance, because this was the only way in practice in which [the architect] intended to get across to [the contractor] the standard required'. That is, cutting corners on specification requires a rigorous (time consuming and expensive) inspection regime. It also reduces certainty. The court didn't comment on the inequity here – without a specification the client and builder cannot know in advance what is acceptable, as this is entirely down to the whim of the inspector. A good specification will be a safer bet all round, though it won't completely remove the need for inspection.

The inspector

The inspector must be competent. The principal of the architectural practice in this case delegated the day-to-day running of this project, including inspection, to an architectural technologist. The court commented 'he was not senior enough' for that responsibility, being neither a partner nor an associate, and was 'overly confident in the standard of work being performed by [the contractor]' because of this. The judge formed the view that the architectural technologist 'was left to run this project on behalf of [the architect] without any proper guidance from [the partner in charge]' – he was 'almost on his own'.

The inspector's lack of experience was a particular problem for the joinery for this project, which was to be of a very high standard. Although it was of a reasonably high quality, there were numerous examples of defective work and poor finish, much failing to meet even an ordinary standard of workmanship. The inspector seemed unable to recognise sub-standard work when he saw it.

Compounding all this, the client 'often changed his mind "when he saw things constructed"', and would himself instruct demolition or rebuilding. The client's 'fixers' likewise issued instructions directly to the contractor, without reference to the architect or QS. Quasi-inspectors of this kind clearly undermine the contractual inspector and lead to confusion.

Written records

The architect kept no minutes, notes or notebooks. There were no records of inspections, and there should have been: 'An inspection can only be effective if the inspector clearly identifies
what defects he has seen and what he wants done about them' and 'the absence of any records
[the architect's] periodic inspections is a matter of particular concern'.

Instructions were issued verbally to the contractor, who then confirmed them in writing. Or not – on a number of occasions (e.g. cupping of floorboards), instructions which the inspector claimed he had issued (let's assume truthfully) had not been confirmed in writing and the contractor had not done the work. As far as the court was concerned, the conclusion had to be that the instructions had not been issued at all. By way of an example, for the unrebated door, 'an ordinarily competent architect ... must take all possible steps to ensure that the defective element is replaced ... he should have written to [the contractor] ...'.

The court explored this issue further with respect to M&E items. Inspectors should, having drawn the contractor's attention to an item of defective work, 'chase the contractor until the necessary remedial work is done', or make a deduction from the interim payments. To do neither is to open oneself up to liability.

Instructions should be issued in writing, rather than verbally. There is no legal distinction between issuing written instructions, and issuing verbal instructions which are confirmed by the contractor in writing. In this case the inspector thought, incorrectly, that he was barred from doing the former so he resorted to the latter, under the mistaken belief that they are not the same thing.

Finally, there was no written criticism or complaint to the contractor. One reason is that minutes of the site meetings were made by the contractor, who was unlikely to record any criticisms or complaints raised at these meetings.

Timing of inspections

From previous cases, and reiterated in this judgement, it is not enough to inspect only at regular site meetings – rather, one must inspect with respect to the nature and progress of the works. In this case, the architect's periodic inspections were only ever at site meetings – they were 'insufficient', 'too infrequent and too rigid'. Other visits were made, but not for inspection, just to 'have a look around'. The judgement made it clear that inspections are not timed just to suit the convenience of the inspector.

Following discussion about alleged pyrites staining of roof slates, the court also made it clear that there is no obligation to regularly reinspect such that deterioration would be seen as it arises. Only two inspections are needed, the initial inspection (when the slates were not stained) and snagging (at which point the staining might have been spotted).

From previous cases, the court made several time-related observations. The inspector can instruct the builder to not cover important work, and inspect later, but it is better to time inspections of important coverable items so as to avoid affecting progress of works.

Work covered but not inspected may not necessarily reflect on the inspectorial regime – this may be acceptable if the work is not important or not anticipated (e.g. outside contractor's program), or if the inspector is confident in the contractor.

Finally, work to be repeated, and therefore important, should be inspected early, to set a benchmark. In this case the inspector set quality benchmarks (e.g. Bedroom 4) too late.

Timely rejection

From previous cases, the inspector should point out to the certifier (e.g. QS) items of defective work, so they can be excluded from certification. In this case the inspector claimed to have advised the QS verbally of defective work, during inspections – but there was no record of this.

The main lesson from this case is that work found to be irreparably defective at a periodic inspection should be rejected there and then, rather than deferring rejection to snagging (which the inspector in this case repeatedly claimed to have done, e.g. for windows with shakes, defective joinery, gaps between stone panels, a gap requiring silicone sealant). Otherwise, the contractor is paid for the defective work and, if he goes bankrupt, the cost of repair must then be borne by the client. The client in effect pays twice. If you like, the inspector's working assumption must be that the contractor could go bust at any time.

So, since the inspector 'should have alerted [the contractor] to the errors in their joinery work long before [the contractor] left site, ... [the architect was] liable for the costs of remedial work if, as
turned out to be the case,
[the contractor was] no longer in a position to carry out the works or pay financial compensation'.

On the other hand it was OK to defer dealing with loose treads in staircases to snagging. Likewise, wall scorching arising from the operation of wall lights, and cabling faults, wouldn't be picked up until snagging/commissioning.

Level of inspection

The court referred to the Ruxley (swimming pool) case regarding liability for the parts of the drainage system that were defective, and for galvanizing to the patio: 'it will always be a matter
of degree as to whether
[a] particular defect should be rectified, regardless of the disruption and cost, or whether, in all the circumstances, it can properly be left as it is'. For inspection, this means that certain non-compliances (e.g. those not affecting the client's 'enjoyment' of the end result, or where the requisite remedial action is out of proportion to the benefit) should be tolerated – blind rejection of all non-compliances will not be supported by the courts.

From previous cases, it is almost inevitable that some defects will not be noticed by the inspector, who is not required 'to go into every matter in detail'. The inspector does not guarantee that inspection 'will reveal or prevent all defective work'. For some defects 'with the best will in the
world, it is much more difficult to say that an ordinarily competent architect should have picked up on the defect in question'

So what are you expected to spot, and what are you not? This case provided quite a few examples.

What you should spot:

  • It was agreed that the projection of some copings did not comply (under 40 mm) and, because this should have spotted in inspections, the architect was in breach of contract.
  • The position of the sarking membrane was changed and drawings issued to show this, including the use of plywood to prevent a water trap. Given the importance of this change to the integrity of the roof, and the lack of clarity in the drawings, the inspector should have checked carefully on site that the plywood was installed as required. The architect did not do this and so was in breach of contract.
  • Periodic inspection should have picked up open joints, etc. in the feature trusses. Again, the architect was liable. The structural engineer was not, even though he had inspected the trusses structurally: 'It was not for [the engineer] to advise as to aesthetics'.
  • Perimeter expansion strips for floor boards were not documented, and this omission should have been identified through inspection – the architect was liable.
  • The architect was held liable for not spotting staining in the external stairs, arising from the contractor's defective detail.
  • The architect should have noted in his periodic inspections that the fire stopping was inadequate between the lift shaft and the garage, and so was liable.
  • The structural engineer should have drawn the client's attention, at the outset of the works, to cracks in the existing granite retaining wall and, even though there was no loss to the client, was in breach of contract.
  • The M&E engineer should have seen, at a periodic inspection, that the hot water return pipe had not been installed, this being an important element of the services. The engineer was liable.

What you are not expected to spot:

  • While there was no dispute that the paint was too thin, the architect was not liable as it was agreed that he could not spot this at periodic inspections, 'unless the inspection took place
    at the time of the application'
    , which was not a reasonable expectation. This logic is one reason that architects don't supervise (see 'Specifying construction processes', NBS Journal 01).
  • Insulation was drawn but not installed – 'it could not possibly be said that this one-off error
    was something which
    [the architect] should have spotted on their periodic inspections'.
  • The inspecting architect cannot 'be expected to identify the particular type of glazing
    compound being used'
  • Since others from the client's side had failed to see the (alleged) pyrites staining of the roof slates, it is unreasonable to expect the inspector to have spotted it. In any case 'an architect's periodic inspections are not intended to (and do not) ensure that the materials supplied by reputable suppliers, and installed by the building contractors, are free from all defects'. This is a very interesting observation. For defects in goods we can rely, for example, on the Sale of Goods Act 1979 externallink.
  • Conversely, for the kitchen, the architect was not in breach of his inspection obligation because 'everyone knew of the deficiencies in the kitchen, and [the client] took an informed decision to continue with the kitchen to completion'.
  • Because the architect was unaware of patch repairs carried out to the rain water tank, and would not have been expected to spot them during periodic inspections, he was not held liable for them.
  • Related to this, the court found that 'defective installation [of the rainwater tank] was not something that [the structural engineer] should necessarily have seen on inspection'.
  • The structural engineer was not expected to make a special visit to site to inspect a couple of gable straps before they were covered up.
  • The fact that the fire alarms were too close to the wall is something that the experts agreed 'might "ideally" have been identified on periodic inspections' by the M&E engineer. Turning on the word 'ideally' the court found that the fact that it wasn't spotted does not make out a case of negligent inspection. This also applied to an extract fan, labelling of a supply air fan in the garage, restricted air discharge to the FCU, incorrect installation of boiler flues, poor installation of sanitary equipment and incomplete installation of an irrigation pump. If the experts hadn't used the word 'ideally' the court might have made a different finding in these cases.
  • Because it is just 'a matter of detail', the M&E engineer 'cannot possibly be criticised for failing to spot the error' in the use of 22 mm supply pipe connections to the bath, instead of the specified 28 mm.

A conclusion is that, if something has been documented or executed, this does not create an absolute obligation to check for compliance. But because failure to check can lead to breach of contract, especially for more significant items (whatever they are), it is perhaps better to inspect too much than too little!


This project didn’t reach the snagging stage. Snagging is limited to 'minor items which can be swept up at the end', e.g. final finishes, but not major items. As already noted, the inspector should not wait until snagging before inspecting and requiring defective work to be remedied. The court gave two principal reasons:

  • Defective work is paid for alongside acceptable work. If the contract terminates, then this amounts to over-payment, as already mentioned.
  • It is cheaper and simpler to remedy defects as they arise, especially if they are in work to be repeated.

In the event of early termination, the court must determine if incomplete work is work in progress (and therefore not ready for inspection, which would be seen as unreasonable and disruptive), or if it should be treated as complete (and therefore ready for inspection and then rectification). Appropriate records, such as adequate photos, would assist in this determination.


While one must beware of imposing an 'unfairly onerous view of the standard to be expected of an ordinarily competent architect', as McGlinn's expert architectural witness sought to do, there is no question that architects and engineers must purposefully inspect the works in progress and deal properly with non-compliance, through written notification and appropriate action.