Your operating result is determined by the “behaviour” of your business processes. Why is the order cycle time for order A so much shorter than that for B? Where are the bottlenecks in my processes? Where does rework occur and what does it look like? What can I do to speed up my cash conversion cycle? And what about compliance?
Axisto answers these and many other questions with the help of process mining. Process mining is a data-driven analysis technique that shows actual process behaviour. We use this technique as an analysis tool for process and performance improvement. And it can either be applied for internal/external audits or be integrated into day-to-day operations to increase insight and support decision-making. Axisto guides you through the maze of different process mining tools, helps you tailor-make a solution that is right for you, and supports you towards a successful implementation.
THIS INTERVIEW WAS PUBLISHED BY THE GUARDIAN
Zoë Corbyn
Sun 6 Jun 2021 09.00 BST
‘AI systems are empowering already powerful institutions – corporations, militaries and police’: Kate Crawford. Photograph: Stephen Oxenbury
The AI researcher on how natural resources and human labour drive machine learning and the regressive stereotypes that are baked into its algorithms
Kate Crawford studies the social and political implications of artificial intelligence. She is a research professor of communication and science and technology studies at the University of Southern California and a senior principal researcher at Microsoft Research. Her new book, Atlas of AI, looks at what it takes to make AI and what’s at stake as it reshapes our world.
You’ve written a book critical of AI but you work for a company that is among the leaders in its deployment. How do you square that circle? I work in the research wing of Microsoft, which is a distinct organisation, separate from product development. Unusually, over its 30-year history, it has hired social scientists to look critically at how technologies are being built. Being on the inside, we are often able to see downsides early before systems are widely deployed. My book did not go through any pre-publication review – Microsoft Research does not require that – and my lab leaders support asking hard questions, even if the answers involve a critical assessment of current technological practices.
What’s the aim of the book? We are commonly presented with this vision of AI that is abstract and immaterial. I wanted to show how AI is made in a wider sense – its natural resource costs, its labour processes, and its classificatory logics. To observe that in action I went to locations including mines to see the extraction necessary from the Earth’s crust and an Amazon fulfilment centre to see the physical and psychological toll on workers of being under an algorithmic management system. My hope is that, by showing how AI systems work – by laying bare the structures of production and the material realities – we will have a more accurate account of the impacts, and it will invite more people into the conversation. These systems are being rolled out across a multitude of sectors without strong regulation, consent or democratic debate.
What should people know about how AI products are made? We aren’t used to thinking about these systems in terms of the environmental costs. But saying, “Hey, Alexa, order me some toilet rolls,” invokes into being this chain of extraction, which goes all around the planet… We’ve got a long way to go before this is green technology. Also, systems might seem automated but when we pull away the curtain we see large amounts of low paid labour, everything from crowd work categorising data to the never-ending toil of shuffling Amazon boxes. AI is neither artificial nor intelligent. It is made from natural resources and it is people who are performing the tasks to make the systems appear autonomous.
Unfortunately the politics of classification has become baked into the substrates of AI
Problems of bias have been well documented in AI technology. Can more data solve that? Bias is too narrow a term for the sorts of problems we’re talking about. Time and again, we see these systems producing errors – women offered less credit by credit-worthiness algorithms, black faces mislabelled – and the response has been: “We just need more data.” But I’ve tried to look at these deeper logics of classification and you start to see forms of discrimination, not just when systems are applied, but in how they are built and trained to see the world. Training datasets used for machine learning software thatcasually categorise people into just one of two genders; that label people according to their skin colour into one of five racial categories, and which attempt, based on how people look, to assign moral or ethical character. The idea that you can make these determinations based on appearance has a dark past and unfortunately the politics of classification has become baked into the substrates of AI.
You single out ImageNet, a large, publicly available training dataset for object recognition… Consisting of around 14m images in more than 20,000 categories, ImageNet is one of the most significant training datasets in the history of machine learning. It is used to test the efficiency of object recognition algorithms. It was launched in 2009 by a set of Stanford researchers who scraped enormous amounts of images from the web and had crowd workers label them according to the nouns from WordNet, a lexical database that was created in the 1980s.
Beginning in 2017, I did a project with artist Trevor Paglen to look at how people were being labelled. We found horrifying classificatory terms that were misogynist, racist, ableist, and judgmental in the extreme. Pictures of people were being matched to words like kleptomaniac, alcoholic, bad person, closet queen, call girl, slut, drug addict and far more I cannot say here. ImageNet has now removed many of the obviously problematic people categories – certainly an improvement – however, the problem persists because these training sets still circulate on torrent sites .
And we could only study ImageNet because it is public. There are huge training datasets held by tech companies that are completely secret. They have pillaged images we have uploaded to photo-sharing services and social media platforms and turned them into private systems.
You debunk the use of AI for emotion recognition but you work for a company that sells AI emotion recognition technology. Should AI be used for emotion detection? The idea that you can see from somebody’s face what they are feeling is deeply flawed. I don’t think that’s possible. I have argued that it is one of the most urgently needed domains for regulation. Most emotion recognition systems today are based on a line of thinking in psychology developed in the 1970s – most notably by Paul Ekman – that says there are six universal emotions that we all show in our faces that can be read using the right techniques. But from the beginning there was pushback and more recent work shows there is no reliable correlation between expressions on the face and what we are actually feeling. And yet we have tech companies saying emotions can be extracted simply by looking at video of people’s faces. We’re even seeing it built into car software systems.
What do you mean when you say we need to focus less on the ethics of AI and more on power? Ethics are necessary, but not sufficient. More helpful are questions such as, who benefits and who is harmed by this AI system? And does it put power in the hands of the already powerful? What we see time and again, from facial recognition to tracking and surveillance in workplaces, is these systems are empowering already powerful institutions – corporations, militaries and police.
What’s needed to make things better? Much stronger regulatory regimes and greater rigour and responsibility around how training datasets are constructed. We also need different voices in these debates – including people who are seeing and living with the downsides of these systems. And we need a renewed politics of refusal that challenges the narrative that just because a technology can be built it should be deployed.
Any optimism? Things are afoot that give me hope. This April, the EU produced the first draft omnibus regulations for AI. Australia has also just released new guidelines for regulating AI. There are holes that need to be patched – but we are now starting to realise that these tools need much stronger guardrails. And giving me as much optimism as the progress on regulation is the work of activists agitating for change.
The AI ethics researcher Timnit Gebru was forced out of Google late last year after executives criticised her research. What’s the future for industry-led critique? Google’s treatment of Timnit has sent shockwaves through both industry and academic circles. The good news is that we haven’t seen silence; instead, Timnit and other powerful voices have continued to speak out and push for a more just approach to designing and deploying technical systems. One key element is to ensure researchers within industry can publish without corporate interference, and to foster the same academic freedom that universities seek to provide.
Atlas of AI by Kate Crawford is published by Yale University Press (£20). To support the Guardian order your copy at guardianbookshop.com. Delivery charges may apply.
Although companies have automated many of their processes, much of the work is still done manually and routinely. For example, employees have to manually enter data and switch between systems, applications and screens in order for the critical business processes to function. Such activities have no added value in themselves. They take time, pose an inherent risk and are expensive.
Axisto replaces these tasks with Robotic Process Automation (RPA), which performs the tasks 24/7, quickly and error-free. In this way, employees can contribute added value by focusing their attention and expertise on those tasks that are important. Freeing up time simplifies the transition from reactive to proactive operational processes. RPA is highly flexible – it can be used in many different areas and can be scaled up and down quickly. RPA adds value by better exploiting and expanding the possibilities of machine learning and artificial intelligence. Axisto provides advice to find your best options and supports you to implement these new ways of working.
CHALLENGE
Maintenance contractor for Rail Infrastructure executes maintenance based on human inspection of rail track videos.
Introduce predictive maintenance technology to streamline and improve inspection.
APPROACH
Gather data on inspections executed to create a large test set of observations.
Develop a deep learning algorithm (based on neural networks) to automate the inspection.
Minimise false positives (judged okay but in fact not okay) and false negatives (judged not okay but in fact okay).
Test and validate the software in the live environment.
Transfer the software to the maintenance organisation.
Are your business processes still up to the required standard?
Particularly now, in the major economic recession that has set in due to the Corona lockdowns, the quality of your business processes have to be top notch. After all, this is the determiner of the operational performance your company delivers.
Over the years your customer base will have changed, but so too has your supplier base, your products and services, your employees, your IT-infrastructure and your organisational structure. With every change your processes have been affected and now they are rather a tangled web with insufficient performance.
Understanding how your processes behave – a prerequisite for improvement
The scale of the current crisis requires companies to move quickly. You want to cut costs, slash cycle time, serve customers faster, get things right first time, improve reliability and be more agile. Therefore, you need a 100% fact-based insight into how your business processes are currently performing – and this can only be obtained through Process Mining.
Process Mining – the gateway to rapid operational performance improvement
How business processes behave is determined by the way those processes are designed and represented in IT-systems on the one hand and how your employees behave on the other. The good news is that this information is already captured in the log data in your IT-systems. This log data can be loaded from your IT-systems into the Process Mining software tool and, literally at the touch of a button, it then displays your process, as it behaves in daily reality, with all the process variants, rework loops, bottlenecks, compliance issues and more. Static, as a process flow, and dynamic, in the form of an animation. In short, a fast and effective way to understand processes and make an effective impact on those things that really matter.
The power of Process Mining
Process Mining (PM) quickly puts an end to “underbelly” discussions because it is fact-based. With PM, both the entire process and an individual case can be analysed – and everything in between. After loading your log data into PM, interactive “deep dives” can be organised with the people involved in the process. Our experience is that the root cause of problems is discovered in a short time and it quickly becomes clear which actions need to be taken to solve problems and seize opportunities.
Steering the company in volatile times
The coronavirus pandemic has caused a swell across business markets, in terms of both volatility and financial risk. To sail the company ship through these rough economic seas, your compass needs to be accurate and the quality of decisions and actions taken must be right. All crew members need the appropriate information to make the right decisions and act accordingly. Unfortunately, this is seldom the case. More often than not the crew sails blindfolded, unable to act effectively; the company drifts off course and targets are not met.
The Performance Management System
This is where the Performance Management System (PMS) comes in. In Part 1 we focused on an appealing vision and ambition – the dot on the horizon to sail to. In Part 2 we considered how to align the cost structure with this vision. In Part 3 we deployed the vision and ambition through all organisational levels and made sure that everyone in the company was aligned and owned the targets. In Part 4 we looked at how to design the key business processes. The PMS ties all of this together.
The PMS is like the control panel in the bridge of a ship. In fact, it is a set of fast and slow Plan-Do-Check-Act cycles. PDCA – the good old Deming cycle. The basic principle is shown below.
On the right is the “P” column: turning the annual budget into a daily activity schedule. The lower horizontal part is the “D” area: the actual business processes where activities are done. On the left is the “C” column: reporting actual performance against the targets. In the middle is the “A” column: the daily, weekly, monthly, quarterly and annual meetings in which actions are formulated and decisions taken to drive actual performance to KPI targets based on the information and insights supplied through the reports. The fast cycles occur shift-to-shift, day-to-day and week-to-week. Above that the slower cycles are active.
It is important to ensure that the PDCA cycles are closed and that the faster and slower ones work in synch. Closed PDCA cycles drive actual performance to KPI targets.
Purpose of the Performance Management System
The purpose of the Performance Management System is to (1) manage the organisation and the business processes to ensure delivery of the company goals, (2) support the development and continuity of effective behaviour of people at all levels in the organisation, and (3) facilitate people to work as “one team – one goal”.
Company:
High-tech equipment manufacturer
Production of low-volume highly complex machines consisting of many subassemblies.
ach machine needs testing prior to acceptance by the client.
The test process is a complex, knowledge-intense process, and testing takes several weeks at both sites.
Challenge:
As the company operates in a market where time-to-market of system enhancements and new system types is critical, the goal was to significantly reduce the test period.
Approach:
Process mining was conducted on a batch of 24 machines.
As the goal was to shorten the test process, the focus of the analysis was on idle times and rework in the process event log.
A performance analysis was conducted to find unnecessary idle times:
Based on the logged test sequences, a process model was constructed automatically that showed how the test process had been executed for these 24 machines.
The resulting visualisation of the rework provided the insights on which the test process was organised differently.
Company:
Contract Manufacturer of durable consumer goods, distributed across Western-Europe
Production volume 2 million units in 24 product families.
Very rigid client quality requirements.
Final product store is a 24/7 operation on weekdays.
Challenge:
Check compliance with formal inventory management procedures and guidelines.
Check compliance with quality assurance procedures.
Check compliance with First In–First Out (FIFO) procedure.
Check work distribution across shifts.
Approach:
The existing Warehouse Management System (WMS) was used to extract data for the analysis.
554,745 events over a 5-month period were included in the analysis.
Together with warehouse representatives the insights generated by Axisto Process Mining® were discussed and conclusions were drawn.
In November 2020, McKinsey published an interesting paper entitled “Value creation in industrials”, a survey of the US industrials sector. The purpose of the analysis was to gain insight into the factors that determine performance in the industrials sector. Value creation was used as an indicator, measured as annual growth of the total shareholder return (TSR). The research covers the period 2014–2019. So, what are the conclusions on how to create value in Industrials?
CONCLUSIONS
The industrials sector is broad and diverse. In order to compare companies in a meaningful way, McKinsey divided the sector into 90 so-called microverticals. More on that later.
The main conclusions about how to create value in Industrials:
Even in good times, TSR performance across and within microverticals is highly variable.
Despite the tailwind or headwind, companies ultimately determine their own destiny.
The TSR performance gap between the best-performing and worst-performing companies within a microvertical is substantial and growing.
Companies with strong balance sheets for 2019 have, on average, outperformed their competitors: the COVID-19 pandemic has widened the gap between the best and worst performers.
Operational performance, and in particular margin improvement, is by far the most important factor in value creation.
HOW CAN WE COMPARE COMPANIES IN SUCH A DIVERSE INDUSTRIAL SECTOR?
While the manufacturing sector performed well at an annual growth rate of 11 per cent between 2014 and 2019, performance varied widely between the ten subsectors. Now the diversity between and within the subsectors is very great. In order to properly identify the factors that determine the performance, the study worked with 90 groups of companies that carry similar products and that focus on a similar end market: the so-called microverticals.
WHICH TRENDS ARE AFFECTING THE MICROVERTICALS?
Five categories emerge from the research: (1) regulation, (2) consumer and socio-economic, (3) technological, (4) environment, and (5) industrial structure and movements of players in the market. Any one of these trends can cause a tailwind or headwind – often both. Measured in revenue and margin growth, these trends predominantly work out well for the top-performing microverticals and negatively for some of the bottom microverticals.
COMPARING MICROVERTICALS AND COMPANIES WITHIN THEM.
First of all, the fact that the company is in a top-performing microvertical is no guarantee that it is a top performer. It can also be seen that the best-performing companies within a microvertical perform substantially better than the worst-performing companies within the same microvertical. The performance gap is substantial and growing.
McKinsey found a 2,600 base point difference in TSR between the best- and worst-performing microverticals. Approximately 30 per cent of companies performed significantly better or worse than what the performance of their microverticals would have predicted. So success depends not only on whether you are in the “right” microvertical; a company’s actions are also important. Individual companies can do a lot to determine their fate, even when headwinds and tailwinds affect microvertical performance. Furthermore, the survey found that, on average, companies with strong balance sheets for 2019 outperformed their competitors, meaning the COVID-19 pandemic has widened the gap between the best and worst performers.
WHAT CAN WE LEARN FROM THE BEST COMPANIES?
To determine which actions matter at a company level, the TSR performance of individual companies was analysed. To this end, the TSR was divided into three broad elements:
1. Operational performance
This element refers to how a company uses its capital to increase revenues and operating margins; this category also includes a company’s ability to generate value for its shareholders in a scenario with no growth and unchanged profitability. The latter is a measure of the starting position of a company.
2. Leverage
Leverage refers to how companies use debt to improve their TSR performance.
3. Multiple expansion
This element refers to opportunities to take advantage of changes in how investors see the future.
Figure 1 provides insight into the way in which companies secured their position.
Figure 1. The way in which companies secured their position.
Of the three elements of TSR, operational performance was found to be the strongest predictor of TSR CAGR from 2014 to 2019 for all quintiles (Figure 2). Operational performance had the highest correlation coefficient with TSR performance, at 50 per cent, followed by leverage (about 30 per cent) and multiple expansion (about 10 per cent).
At the top-performing companies, operating performance contributed to 18 percentage points of the 27 per cent TSR growth. And for the worst performing companies −6 percentage points of −11 per cent TSR growth.
Figure 2. Operating performance had the strongest correlation with the company TSR.
Within the operational measures, margin expansion was a major contributing factor and also the strongest determinant of the company’s TSR performance (Figure 3). With a 90 per cent correlation to business performance, the profitability extension (margin) adds an average of 8 percentage points to the 18 per cent operational performance of the top performing companies and takes 8 percentage points away from the lowest quintile companies, where the business performance is on average −6 percent.
Figure 3. From the operational statistics, margin expansion proved (often made possible by technology) the main determining factor for the company’s TSR.
Looking at the top-performing companies, it turned out that their success had depended mainly on taking three steps:
HOW DO YOU ENABLE SUCCESS AND HOW DO YOU MAINTAIN IT?
To further increase the likelihood of continued success, companies need good supervision. Executives must balance their time between creating and executing strategies, and periodically reassessing and rebalancing the business portfolio. Along the way, they should look for ways to improve earning power through rapid (two-year) cycles of margin transformation, leveraging technology wherever possible.
Industry 4.0 is in the spotlight. And rightly so. The possibilities are great: higher productivity, a better customer experience, lower costs and perhaps a new business strategy with innovative products and services. And there is an outright need: without Industry 4.0 a company has a limited future. Unfortunately, many Industry 4.0 implementations get stuck. Let’s find out why this happens and how to prevent it happening to you.
DATA
There can be three issues with data: not good, not available, poor quality. This is often due to IT systems not being set up properly, data not being entered or being entered incorrectly, log switches to register log data not being set correctly, or the data entered being of poor quality.
In addition, the knowledge of business processes is seldom up to standard. How do processes behave in daily practice? How should they run? This means that people are unclear as to which data should be captured and how the data should be managed.
It is therefore important to know the business processes and how they work both in theory and in practice. This is the basis for a good KPI and reporting structure. Getting this right will ensure clarity around which data must be collected, which information is required for whom at what time and how to manage the processes for maximum effect. It will also mean that data availability and quality will increase – thus building the foundation for Industry 4.0.
ORGANISATIONAL SILOS
Many companies still have a strong departmental orientation instead of an end-to-end process focus. This leads to limited insight into and understanding of the interdependencies between functions and departments. A strong departmental orientation also means that data is locked up in silos.
Industry 4.0 focuses on the integrated control of the end-to-end processes that run through various departments and even across company boundaries. That is why departments are asked to work together seamlessly and to share data and information. An effective IT infrastructure facilitates this.
CAPABILITIES TO COLLECT AND USE DATA
The introduction of Industry 4.0 requires a significantly higher level of knowledge of the
industry, of business processes and of analysis applications. At every level in the company and within every position, people must be able to handle data well and be skilled in its analysis.
The technical structure of these cyber-physical systems is becoming more complex, and more and more decisions are being made by algorithms. Therefore, it is important that companies develop the knowledge and skills to build applications and assess the behaviour of algorithms and the insights they provide. The introduction of Industry 4.0 requires intensive collaboration between departments and disciplines to develop people and resources at pace.
VISION AND ORGANISATIONAL ALIGNMENT
The introduction of Industry 4.0 affects all aspects of an operating model. The top team needs a shared vision about the value that is required for various stakeholders, and how that value is delivered – the operating model.
Too often, a joint vision is ill-considered and not adequately thought through, resulting in insufficient alignment with the roadmap. In such a situation, an implementation inevitably comes to a standstill.
THE HUMAN FACTOR
The biggest challenge in an Industry 4.0 implementation is not so much choosing the right technology, but dealing with the absence of a data-based and digital performance culture and the corresponding skills gap in the organisation. Investing in the right technologies is important – but success or failure ultimately does not depend on specific sensors, algorithms or analysis programs. The crux lies in a wide range of people-oriented factors.
Since Industry 4.0 transcends not only internal departments but also the boundaries of the company, its success is predominantly dependent on skillful change management.
CONCLUSION
In essence, the reasons why Industry 4.0 implementations get stuck are no different than with other company-wide transformations whose aim is to create a sustainably high-performing organisation. It will not surprise you that the chance of failure is roughly the same: 70%.
Therefore, in the first instance, do not focus too much on just the technical side of the transformation. Instead, concentrate on skilful change management. The technological content side of the transformation is not your main problem. The development of a data-based and digital performance culture and the corresponding skills set is.
