Though associated with the RAF for most of his life, Phil Byrom had a private sector IT career before joining the MOD early in 2009. He took up a logistics support post on promotion later that year, during which he was fortunate to attend a number of Staff College courses that improved both his strategic knowledge and academic rigour, allowing him to once again move on promotion into his present Personnel Strategy role within HQ Air Command. A passionate advocate for innovation, he represents the RAF in numerous forward-looking forums and has authored papers for the Defence ‘think tank’ the Development, Concepts and Doctrine Centre (DCDC). He is particularly interested in Artificial Intelligence (AI), data analytics and automation, while also dabbling in behavioural psychology.
Those familiar with this competition will know what pedigree comes with any submission by Phil, who has won the competition more than once in the past. This essay took a well-structured and strategic, trilateral approach to people, policy and technology which had one marker “hooked from beginning to end”. What was compelling in this submission was the focus on the people element, which neatly and accurately reflected if not the content, certainly the message conveyed by the majority of the other entries – the need for leadership to sometimes lift its head, take a longer view and identify early what threats and opportunities there are for the whole force approach with the advent of burgeoning technology and change. Captivating in both style and content, this essay is entirely deserving of its place and reward, as runner up in the 2018 competition.
The content and any views expressed in this essay are a personal perspective, reflect the personal views of the author, and may not necessarily reflect that of UK Defence.
We live in interesting and exciting times, with myriad innovations proliferating and the pace of technological development apparently ever-accelerating. Nowhere is this more evident than in the logistics domain. From warehouse automation and Artificial Intelligence (AI) supply chain optimisation to additive manufacturing and synthetic biology, the opportunities are boundless. As the incoming Head of Defence Logistics, I should immediately recognise that the Defence Support Network (DSN) transformation programme can only begin to scratch this ‘opportunity surface’ and must serve purely as a springboard to greater things. It must be seen as only the first iteration in a cycle of incremental enhancements, whose spiral development progressively evolves our technological logistical capabilities through the fusion of Continuous Improvement (CI) for evolutionary marginal gains and revolutionary step-changes through the Research and Development (R&D) of emergent Science and Technology (S&T) advances.
It has, however, long been recognised that any successful transformation programme tends to encompass a ‘golden triangle’ of which technology is but one element. The remaining sides are process and people, this last being arguably the most important of the triad, yet paradoxically also that which is most likely to be overlooked or afforded only cursory attention amounting to little more than lip service when transformation is in the air. Therefore, my priority when taking post as Head of Defence Logistics would be to address the processes and people involved in logistics, to ensure that they are afforded commensurate weight compared with technology in any transformation programmes. Indeed, if anything, I should prioritise people over the other two sides, as the implementation of both process and technology stands or falls entirely upon the aptitudes and attitudes of the people entrusted with taking them forward.
The Golden Triangle
This essay will examine each side of this ‘golden triangle’ in turn, in what might be considered increasing order of importance, and discuss related initiatives that I would wish to invoke. The overall intent is to take an integrated approach across the full spectrum of logistic capabilities, whereby the 3 elements of people, process and technology are complementary and combine seamlessly to become each other’s force multipliers. Its inescapable conclusion is that, rather contrary to conventional wisdom and customary practice, it is the people element that proves most pivotal and on which greatest emphasis should be placed in order to deliver the desired effects and outcomes.
We begin with what is surely the most familiar, but also the most seductive, of the triumvirate, technology. It is often easier to visualise the potential of new tools and equipment than it is to realise this potential through effective implementation. However, that is not to say that there is no merit in pursuing innovative technological solutions. On the contrary, recent developments could greatly improve our logistical capabilities and a number of initiatives should be invoked or sustained in the cause of delivering such improvements. For example, research facilitated by the Defence and Security Accelerator (DASA) into autonomous last mile resupply will need maturation and exploitation to fulfil its potential, whether separately or subsumed as a strand within the DSN programme.
Autonomy – much more than robotics
Autonomy in general is an obvious, but nonetheless important, technological facet which could be the subject of one or more initiatives. More than one initiative might be required, since autonomy in all its forms promises to be so ubiquitous as to extend far beyond the logistics arena, cutting right across Defence, Government and indeed society in ways that massively complicate and may well preclude independent development of logistical applications at scale. Toward the left of arc are specific logistics applications, such as the collaborative robots developed by Ocado to roam and manipulate the ‘hives’ within their automated warehouses. While innovative, their technological bases are well founded and proven. Their infrastructure is scalable and modular, raising the possibility of deployable modules for Defence use. This is what most people picture when they think of autonomy in logistics, but robots and other physical manifestations are only part of the autonomy landscape, the far larger opportunity being in back office systems remote from the front line, yet directly relevant to operational success.
Opportunities exist to automate numerous processes currently reliant on human intervention, in logistics among many other disciplines. From the simple and increasingly industry-standard Robotic Process Automation (RPA) tools to more advanced applications that harness Machine Learning (ML) and wider AI, less and less human intervention will be needed to enable future supply chains. RPA overcomes system integration limitations, such as security constraints or the lack of Application Programming Interfaces (APIs) by directly replacing the user in interactions via the standard user interface. For example, RPA can replace a human who spends all their day re-keying data between systems that cannot interface directly, or copying and pasting details from emails, spreadsheets or web pages into corporate applications, without needing anything more technical than does the human user, such as login credentials and a simple workflow. It cannot only perform the same task without the need for a human, it has the potential to do so faster, more repeatedly and reliably, and consequently with fewer errors than the human user. This in turn promises to improve the completeness, reliability and consistency of data input to the system, with beneficial side effects for subsequent analytics and decision support.
Artificial Intelligence (AI)
AI could benefit logistics in various ways, from optimising supply chains and resupply routes to avoid hotspots and enhance security by reducing the predictability of supply lines, to analysing stocks to improve the efficiency of inventory management or automating the quality assurance and standards compliance of new and existing suppliers and items. While current AI is likely to fall into the ‘narrow’ category, being specific to a particular task, the basic models can be used as the basis for more than one application, being retuned and retrained to work with a number of specific datasets for specialised purposes.
For example, an AI model may be trained and tuned to monitor the fuel consumption of road vehicles, making predictions and either recommending efficiencies or even enacting them, if this is appropriate in the prevailing operating environment. Generalising this use case a little suggests the potential to tune a similar model for fuel monitoring on other vehicles, such as aircraft and ships. However, generalising it slightly further raises the possibility of monitoring the use of other resources, whether the consumption of other forms of energy, like electricity, that of other consumables, whether goods like food, water and spare parts or services, or the production and disposal or recycling of waste products and surplus commodities.
This opens up a raft of logistical applications, from predictive analytics to optimise equipment maintenance, moving away from time- and condition-based maintenance or replacement and toward specific life expectancy predictions for each individual component based upon its use in context, to automatic reordering and replenishment of consumables within smart buildings, or even smart bases, where resource usage is continually monitored via arrays of Internet of Things (IoT) sensors feeding data points into the AI model. Extrapolating to its logical extent, this also enables smart working environments, whereby aspects like temperature and lighting can be monitored and controlled by AI via feedback loops, improving the satisfaction, comfort and, consequently, the productivity and engagement of personnel, while concurrently reducing the load on human facilities managers.
Other technological solutions can support and enable these kinds of automation and ought to be considered as part of any integrated technological initiative. Innovations like RFID and QR codes to passively tag and track inventory items and consignments, environmental sensors to monitor the conditions in which items are stored so as to actively manage their shelf lives, 4G networks to control collaborative autonomous vehicles, robots capable of repairing themselves or each other, and additive manufacturing to generate rarely needed yet critical stock items on demand, are all emergent technologies that should be woven into an integrated future logistics landscape.
Underpinning much of the above is the fundamental pre-requisite of effective data fusion. The volume and velocity of the data that we do and can expect to collect in future is enormous and its veracity may be questionable for a variety of reasons. Moreover, separate data sources are of limited use unless and until they are brought together and a view can be taken across them all, creating a richer and more contextual picture. Doing this effectively and cleansing the data such that it is good enough for the specific purposes required enables predictive analytics that can further streamline the logistics chain.
If this sounds simple and trivial, it isn’t. There is a maxim that Data Scientists spend 80-90% of their time cleansing and preparing data, compared to only 10% spent on Data Science proper, (including complaining about the amount of time they spend on cleansing and preparing data!) This has been identified as a key requirement for Defence, which is now funding research into the use of AI for data cleansing through its strategic relationship with The Alan Turing Institute and the maturation and exploitation of the outputs of this research is one initiative that should definitely be taken forward in the logistics field. Clearly, this should also consider advances in computing that increase the processing power available to AI and to Data Science algorithms, particularly the advent of quantum computing, which has the potential to overcome limitations of conventional architectures that might otherwise limit the continued advance of Moore’s Law.
Of course, automation alone does not guarantee success. Automating a bad process doesn’t improve the process, it simply enacts it more quickly and consistently. At best, this is likely to be inefficient. At worst, it might degrade the situation, were increased speed to amplify issues inherent in the process and exacerbate them, either directly or via dependencies elsewhere in the process chain. That said, data analytics and AI can shed new light on the efficacy and the efficiency of extant processes, enabling their iterative refinement for peak performance. This brings us neatly to the 2nd side of the ‘golden triangle’: process.
Process is important for a number of reasons. First, as we have already identified, it benefits us nothing if we enshrine poor processes in new technologies. Before or during the analysis of opportunities for automation and other technological aids, we should examine existing process and attempt to refine it into its most efficient and effective form. Of course, in an ideal world, it would already be in this form through the practice of CI. However, in the real world, there will be few, if any, optimal processes that are as good as they could possibly be in the context of human-centric operations – and potentially fewer that are optimised for new technologies like automation and AI. This is because new techniques introduce new ways of working that offer opportunities for increased efficiency if the process is designed with them in mind and to take best advantage of their strengths. For this reason, careful analysis should be conducted into existing processes in light of any proposed technological changes, with the aim of optimising them to exploit such changes to greatest effect.
Prove/improve process before implementing new technology
For example, a number of inefficiencies and indeed deficiencies came to light in the logistics support to the RAF Tornado force. As the fleet was being run down, spares were recovered from a large number of airframes, but their management was poorly controlled, resulting in wastage and unnecessary expense in maintaining stock levels that were larger than required or parts that were otherwise surplus to requirements. Meanwhile, the inquiry into an incident at Newcastle Airport initiated a deep dive into the Tornado engine overhaul facility and identified serious deviations from established engineering practice that led directly to the incident under investigation. In the first case, the process was inadequate to the task, whereas in the second, the process was sufficient but poorly executed, supervised and audited. Either problem would need to be analysed and addressed before attempting to embark upon the introduction of new technologies, otherwise all we might achieve is to teach our Artificial Intelligence to act in as lax a manner as the human intelligence that preceded it.
However, such analysis should go deeper in looking for opportunities to improve the process yet further in consequence of any proposed technological advances. For instance, would it still be necessary to have a supervisor check the work of a junior technician, then to have a senior colleague double-check their work, if the technician were to be guided by Augmented Reality (AR) which not only shows them and walks them through the relevant process, but can also provide them with instant feedback upon the quality of their work and its compliance with the appropriate standards? There is a trust element to this, whereby those responsible for safety and assurance might not feel immediately comfortable with the idea of leaving such checks to the machine but, as confidence builds over time, the efficacy of such systems and the sense of reliance upon them should become increasingly clear and this likelihood should be factored into the reshaping of processes from the outset, to avoid the need for significant subsequent re-engineering.
It should be recognised that the effects on process are not necessarily confined to the directly related processes and that there might be 2nd, 3rd or nth order effects across the wider Defence Enterprise. Whilst it might not be feasible to conduct a fully holistic system of systems audit of the entire organisation, nonetheless there ought to be some form of root cause analysis of the dependency tree to ensure, as far as possible, that every facet of each particular process has been identified and analysed for potential improvements. The prime example is in the training environment, whereby not only the content of training courses might need to change, but also their nature, delivery and scope. Taking the above example, it is necessary to teach students how to use the AR capability and the implications for supervisory assurance, which may lend itself to completely different delivery models to those currently used. Moreover, if a student no longer needs to hold so much detailed information on process and equipment in their memory, because AR will guide them as they go, perhaps training courses can be truncated in content, length or complexity, generating efficiencies of cost, time and instructional capacity. Allied with concurrent technological advances in training, for example virtual instructors and bitesize self-learning modules, there is transformational potential throughout the training pipeline. Another initiative that I would invoke is a thorough review of logistics processes as they are subject to technological change, ensuring that each review fully explores the process dependencies and seeks to exploit opportunities for improvement across all affected touchpoints.
Pan-Defence process partnering
Of course, we ought not to lose sight of the fact that many Defence logistics personnel are not Service Personnel or even Crown Servants, but contractors employed by industry partners in a Whole Force construct. Indeed, the Tornado Force examples cited above involved services provided under contract and so we must ensure engagement across all partner organisations with the process of reviewing and revising processes in a truly collaborative spirit. That is not to suggest that Industry is unwilling to change or to partner effectively, but sometimes there is a tendency to regard contracted services as a ‘black box’ or closed book into which customers have neither need nor right to enquire. This can be dangerous, particularly in places where the now outdated process might have been inherited from Service sources many years ago and is perhaps managed by ex-Service personnel whose frame of reference is of the same epoch as the outdated process and who are unaware of subsequent developments. This need not be a problem, but it does highlight the importance of effective partnering whereby customer input is welcome and encouraged in the interests of best practice, with reciprocal engagement where the Contractor fully supports both the design and implementation of improved processes that leverage the introduction of technological enhancements. Such discussion of the Whole Force brings us to the 3rd, final and arguably the most important side of the ‘golden triangle’: people.
It has often been said that people are our greatest asset. However, they can be expensive, a fact of which Defence people are becoming increasingly aware in an efficiency culture that is perceived, rightly or wrongly, as being all about cuts and savings. This is especially true when transparency dictates that what the Government spends, including on Defence, stands open to public scrutiny. It is inescapable that personnel costs comprise a sizeable chunk of the budget. This is not solely a military issue. The Whole Force paradigm allows some flexibility in personnel type, and consequently through life cost, but creates its own challenges, with Civil Service headcount caps or cuts and tension between the staffing and service levels we expect and those that we can afford to sustain. Furthermore, legal precedents are being established in cases involving Government contractors with precedents that make organisations responsible for their workers, even if those workers are not directly their employees. This adds a whole new moral and ethical dimension to the employment of contractors, over and above the duty of care with which we are so familiar.
Perceptions of personal threat and resistance to innovation
If we now wish to overlay not simply major change programmes, but fundamental shifts in our assumptions about who works and how, it is not difficult to envisage a scenario in which trust is severely eroded and people interpret transformation and technological advances as simply cuts by another name. In fact, we may already be there, as many people are already scared of technologies like automation and AI, which they feel sure will take their jobs, leaving them on the employment scrapheap. There is talk of ‘hollowing out’ middle management, leaving only low skilled, low paid jobs and highly professional, creative jobs, with the remainder providing rich pickings for robots and automatons. Whatever the real automation gradient, combine this with the prospect of Defence efficiencies and Civil Service reductions and it is easy to imagine unrest increasing across the Whole Force and resistance to change developing that threatens to rival the luddites in its destructive intent.
In fact, the entire and wholly laudable Defence innovation agenda may founder on the people element due to obstruction not only from those who feel threatened by the relentless march of progress, but also from those who dislike change or see it as being change for change’s sake, and those who simply fail to see the vision or grasp the need. The biggest threat to success for Defence innovation is, however, the number of people who are so busy and so focused on delivering today’s effects that they cannot raise their gaze above the immediate and take the long view.
That is certainly not to say that the Defence Enterprise is devoid of innovation. Similarly, there is little truth in the assumption that innovation is vested solely in the younger generations. Yet, being a staunchly hierarchical and rank-based organisation, Defence suffers from the problem of ‘blockers’. Anyone with a safety background is likely to recognise the Swiss Cheese model, whereby risks are like holes in slices of gruyere, but only when all the holes align does the risk materialise into an incident. Blockers are the malignant reverse situation, where good ideas cannot get through the organisation to where they can be exploited to greatest benefit as their path through successive Swiss Cheese slices is frequently blocked at one level and proceeds no further. Sometimes, enterprising innovators can circumvent blockers by finding another route whose holes align, but often this is impossible, especially if they are at a junior level and have no senior network on which to call. Even when good ideas do make it through, there can often be ‘Chinese Whispers’ to contend with, whereby the intent is occluded by translation at multiple organisational layers, or where existing supposed Subject Matters Experts (SMEs) corrupt or water down an innovative vision into something more conventional.
So, the final and most important initiative that I would launch as Head of Defence Logistics is a People one, to flatten the organisation and empower those within to not only originate, but also exploit innovation. We know that people yearn for Autonomy, Mastery and Purpose and the innovation agenda is an ideal time to demonstrate that we trust our people, empowering and resourcing them to harness their ingenuity. It has never been more important to take the long view, learn from past triumphs and failures, and ensure that we leverage the doctrine of Mission Command as effectively across the full spectrum of Defence Logistics as we do when on operations. I believe that Sir Arthur ‘Bomber’ Harris had the maxim that everyone under his command was empowered to say ‘yes’, but that only he could say ‘no’. I think we might do a lot worse than to adopt this premise and let a thousand (innovative) flowers bloom.
In summary, I am minded to para phrase Corinthians 13:13, ‘And now abideth people, process, technology, these three; but the greatest of these is people’.
This essay was edited for publication by Chris Markey FCILT, Chairman CILT Defence Forum.
  The Defence Support Network is an evolution of the UK MOD’s Joint Supply Chain and recognises the increasing complexity of logistics, information, and enabling activities for defence on operations, and preparing force elements for operations. It includes Regular and Reserve Forces, other government departments, allies and industrial partners.
 DASA is part of the Defence Science and Technology Laboratory and the MOD. Devised as a conduit for collaboration between industry, government defence and security departments, academia, and allies, the Accelerator should allow the rapid development of innovative solutions to the most pressing security challenges.
 The whole force approach is one that seeks to balance reserves, regulars and contractors, forging them into a coherent whole that delivers the required capability at a lower cost than if it were a wholly regular military affair. [This is not necessarily cost-driven but may be simply a matter of a lack of in-Service skills and so is Suitably Qualified and Experienced Personnel (SQEP) driven}