There have been several practical problems with the APS over time. They are not always about a "lack of innovation" or some other buzzword or flavour of the month. There are some extremely capable people across the service, but often their message gets lost. Further, things take far longer than they should to set right much of the time. I put much of this down to a few main reasons, and am hoping that the review finds effective solutions to these practical problems. These factors are not time specific, so addressing them would setup the APS well over time (not just to 2030 or whatever you have specified).
1) Clearance processes
Unfotunately a film crew is not needed for the Utopia TV show, you could literally take cameras into offices to make the show! Often for simple briefings of a page or less, up to 10 levels of clearance may be required, sometimes more. Then someone senior goes on leave, the document has unhelpful contributions by an acting person, and then is re-revised when the original person returns from leave. By that time, the document may have been redrafted 20 times, 30 times, often with one middle manager having their suggestions undone and redone at different levels. Locally, work areas may try and address this inefficiency through different means, but these prove to be universally ineffective and a more systematic approach/change is needed for the APS to be agile into the future. In one agency I have worked for, a facebook post had to be vetted by 7 different levels of management. Often by the time they were vetted and placed online, they were obsolete (which is completely unsurprising). One could blame the structures within the agency, but again across the service these issues persist. Additionally, the very senior managers instituting these sorts of processes are not the ones navigating their way through them, so they do not have visibility over the problem (they may well be blind to the problem). Further, there is no evidence to suggest that the quality of briefings increases based on these obtuse clearance processes, in fact it is known that several briefings and important public documents still regularly contain significant mistakes. Another consequence of this is that by the time is document is prepared it is stripped of all meaning - obviously preaching about plain English is different to actually using it. Note that I am not suggesting against a sensible QA regime; it is important not to polarise arguments for the purpose of dismissing their merit.
This is a system problem possibly brought on by excessive hierarchy and generalisation but with little regard for specialisation and specific employee expertise. About hierarchy, senior managers are often extremely busy. But they are made extremely busy by much of this type of work. Yet there is a self-fulfilling component of this, where they are the very people who could choose to make themselves less busy by deciding on better processes. But they do not - whether it be for reasons of empire building, self-importance or whatever. An important component of good management is listening to the experts and allowing the right people to do work efficiently, yet this is rarely achieved in the APS. Managers seem to feel the need to "value add" somehow, all the while the many levels are actually obstructing progress and making the simplest tasks interminably slow. Given that managers may have selfish motivations to create obstructive bureaucracy, a more risk based and balanced approach to QA needs to be embedded in a future APS structure. In terms of generalisation of skills compared with specialisation, I will touch on this later.
2) The in-practice remit of the APS to provide robust advice and service delivery
The APS is over time becoming too close to government, with very senior management transitioning to yes men in unbelievably political circumstances. Politicking should not be the role of the APS, which should instead implement government directions in the best possible way and provide the best possible informed advice to government. For example senior APS management should not need to attend extremely expensive training courses in how to avoid embarrassing the government at senate estimates, they should appear and answer questions fully and frankly. They should do the rest of their job the same way, as should the APS broadly.
One secretary of a government department, who is known to be extremely competent, kind and capable, once mentioned the need to "manage the risk of FOI." At officer level, I have heard conversations at my workplace about the writing of executive minutes which deliberately omitted information in case of FOI (this would be extremely common, against the public interest and in my opinion appalling behaviour from all involved though I won't name names). This is a great illustration of the current problematic secretive environment shrouding the APS. It should be possible to have a department provide advice to government knowing the government may not agree with it. This advice should not be watered down or hidden from public view, but in some circumstances it could be made clear the the advice is about the best way of implementing a government policy. The government would then, as the decision makers, then be at liberty to disgregard the advice for justifications of its own description. This is how a healthy public service works, not as a defacto branch of the executive as it has mostly become. Just to be certain, FOI is not the problem nor the cause, the relationship between the government and the APS is.
One way to implement these types of changes would be to legislate the independence of a much larger of the APS (there may need to be exceptions for some situations e.g. during workplace bargaining. For example, the Australian Bureau of Statistics operates independently from government through a legislated mandate. Although the defined accountability of the bureau to government may not have been sufficient in the context of Censusfail, this is a scenario which I am confident could be avoided through well written legislation. In general I am sure a better balance could and should be struck.
3) Merit based staffing
Despite the processes around merit based recruitment, it is alarming how easily people can hire their mates or favourites. For example they can write selection criteria to suit not the role, but the applicant they prefer. This issue is perpetuated by generic ILS criteria which do not include job specific skills. In some very limited contexts, a "job specific criteria" may be one of the six or seven selection criteria, and is rarely weighted at 50% or more like it should be. For any given job, there should be a job specific selection criteria (matching the prescribed role) which should be weighted to at least 50% of the total. How can I forecast the weather if I can "achieve results", "liaise with stakeholders" and "develop project management frameworks" but I cannot perform the required tasks to forecast the weather?? I am aware of so many cases of capable people being denied opportunities in the APS with feedback such as "you're not good enough at strategic thinking" provided, which is such an empty platitude devoid of practical meaning that it wastes capability and fosters favouritism. Increasingly as policy problems and service delivery become more complex, specific skills are needed to perform these tasks, and worryingly much of the public service is geenralising. The results (Censusfail and other "tech wrecks", purchases of defective defence equipment etc) are obvious.
A further disturbing development in staffing is the inappropriate use of contracting or procurement. I have no ideological position against contracting etc, however it is inappropriate for defacto full time permanent long tenured staff, as it avoids merit principles. The Canberra Times wrote an excellent article about it:
https://www.canberratimes.com.au/public-service/job-for-the-boy-the-public-service-is-giving-up-on-the-merit-principle-well-all-regret-it-20171204-gzyg75.html
This contract was not even widely open for applications, but only "available to one". This reeks of cronyism, and despite the department's response, they never stated that they ran a merit based recruitment process, only procurement processes for an activity which is not fundamentally a procurement. Further it is likely in the modern day APS that a major contributor to gender pay disparities is cronyism.