The theme of 'don't fear the algorithm but don't believe it either' was pursued at ThinkFM's session on disruptive technology, reports Martin Read.
05 June 2018 | Martin Read
Such is the relentless march of technology that in 2018 most if not all business sectors are evaluating the extent to which data analysis is likely to form part of, or even entirely supplant, existing roles.
Yet facilities management can perhaps claim even greater potential change, citing the scale by which sensor-enabled measurement of both equipment and personnel could reshape the FM service landscape.
Here, suggested several ThinkFM speakers, is the potential for a disruption jamboree: Massive data volumes made absurdly easy to access, interrogate and act upon - with time management, process structure and value use-case the principle variables. Consultant Lionel Prodgers spoke, for example, with great enthusiasm about the scale of the opportunity our sensor-led future will provide.
And yet disruptive tech must first be parsed through just as much disruptive thought, suggested speakers Colleen Conklin (Sodexo) and David D'Souza (Chartered Institute of Personnel and Development), each making compelling cases for considering the emotional and human impact of a workplace world purely characterised by micro-measurement.
Step back from the fray, they argued; consider the bigger performance picture. Productivity is as multifaceted and difficult to enhance as the wider economy; assumptions that sensor-enabled management is a productivity panacea in and of itself should be robustly challenged. Over the next five pages we frame the key issues that came out of the session.
Former BIFM chair and Agents4RM managing director Lionel Prodgers introduced ThinkFM's 'Preparing for the revolution' programme covering big data, artificial intelligence and automation. The issue, he said, is how best to deal with an informational 'tsunami'.
"None of us could have predicted five years ago what we can now do," Prodgers continued. "We have a plethora of technologies all processing information."
Prodgers slipped a multiple-sensor device from his pocket, explaining how it can display screen ambient temperature, humidity, lumen levels, noise levels all from on a phone screen; a world of instant micro-management seen from anywhere in the world.
A key issue is in determining who is responsible in the outsourcing process for retaining and managing data? "We need policies for who owns the data, where it resides, and who has the responsibility for keeping it up to date."
On FM's role in BIM
"Many of you will have faced the conundrum of what to do with Building Information Modelling (BIM) data and how we relate it to computer-aided FM (CAFM) and other systems - but it's no longer the challenge it once was."
A new generation of integrated servers is helping to ease the issue of parsing data from BIM systems into others through the non-proprietary Construction Operations Building Information Exchange (COBie) data format.
"We can now manage the use of a specific asset against design criteria by direct reference to manufacturer's data."
On using manufacturers' data
FMs should make more of the direct links to component manufacturers' databases that sensors now allow.
"We are talking about expanding information related to a particular asset, not by taking more data into proprietary systems but by going out to manufacturers to get their data, in real time, on the asset we are talking about.
"What is required is thinking about how we can thread this data together and how we use it. We don't have to spend a lot of money to make this happen.
"A lot of functionality is being delivered by systems we already know - Microsoft, Google, etc. - using their analytics and cloud software on our own PCs and devices.
"We can do more with asset attributes once we've linked the asset to the manufacturer's site, where technical data such as installations, maintenance manuals and everything related to that asset is accessible via URL links and joined with downloadable standards such as SFG20. It doesn't require surveyors and technicians to go out surveying for the asset register from scratch."
On our newly tactile relationship with data
"I like the idea that data is becoming tactile," said Prodgers. "We can now touch it and move it around with our hands; less effort is going into the management and visualisation of data."
On establishing 'absent' data
What happens when asset data is unavailable? Prodgers spoke of putting asset registers together from scratch in the Middle East, where there can often be very little accurate information.
"In some cases, to create an asset register you'd need hundreds of surveyors working for months - and there'd be no guarantee it could then be kept up to date."
The Saudi Arabia department of education faced having no data on the assets in 30,000 schools. The solution? Create basic online pro forma lists and ask those who really know, such as school principals, to count the fire extinguishers, air con units, broken doors, etc. No specialists needed, and the form is interactive and accessible to those who need them. In Saudi Arabia, existing schools are in this way feeding useful information into the ministry tasked with building new schools. In essence, the fast-tracking of data capture through those on the ground is influencing massive capex decisions.
The lesson? "We don't have to invest millions in new systems because apps that allow us to interrogate the intelligence that already exists. We shouldn't allow ourselves to be steered by technological innovation alone."
Guillaume van Eeckhout of Siemens brought together the technologies already changing the workplace management calculus:
Voice activation, hand gestures
"Your building is increasingly used as an individual interface, with screens strategically positioned and responding with information as you walk past."
"Millennials are driving change. The importance of where, how and who you work for is changing. Culture match is more important. What does your workspace say about your company and how can you use that to attract and retain talent? As our customers move towards these kind of KPIs, we have to move with them from building insights to user insights, moving from facilities centric data sets to user-centric data sets."
"You can make a building as efficient as you want - it stops being efficient the minute someone walks through the door. People in buildings are disrupting the way you set up your BMS, the way you set up your thermostats "
As buildings become more autonomous, the challenge is how we make the building respond to occupant stimuli, not just make its own decisions. It's a conversation about automation vs user-centricity.
Amazon Go tracks you and your grocery choices as patterns and preferences. That's massive automation and user-centricity at the same time.
"In the real estate world, WeWork is pushing the limits in terms of extreme flexibility in leases, work space, customising comfort and access and developing a large community of WeWorkers. This will impact teams or clients directly."
Acting as the antidote/palate cleanser following a raft of sensor and automation based presentations, CIPD's David D'Souza's presentation was well received.
AI and automation - meeting it in the middle
Too much talk of artificial intelligence and automation risks taking the 'human' out of workplace futures.
D'Souza spoke of a range of possible futures, from the full Star Trek of a life of total ease through automation to the nightmare triggering of the Terminator 'Skynet' and robots killing us all.
"The bit I'm really interested is in the middle, where whatever your view there is going to be bridging period where we see rapid changes in the nature of the workforce.
Implications of new data scepticism
"One thing HR has in common with FM is that we've spent last five years on grand data binge, where if we can measure it we're going to find a way to measure it, manage it, make it predictive and do everything possible with this data.
"If we become more sceptical about data (D'Souza spoke of Facebook, Edward Snowdon and the recent Cambridge Analytica scandal), supposedly 'smart' devices are then seen as 'stalky' devices' - and that has implications for many freshly minted means of measurement.
"Biometrics: two years ago you'd be incredibly excited about the concept of giving everyone fitbits and tracking their activity around the office, seeing how that impacts their well-being. But now? It feels like iffy ground, doesn't it?"
Asymmetrical economic impact
Estimates about the impact of automation on businesses range so wildly that they're in danger of parody, "but assuming we do fundamentally automate, productivity figures will shoot through the roof. Yet think this through: if you measure productivity in terms of output per worker per hour, then one robot equals infinite productivity - and no one is earning".
This point is worth extrapolating, he says. "The economy is a complex, cyclical ecosystem. So if you automate retail jobs in bulk then there's less disposable income in the economy. More aggregated wealth for certain people, yes - but you don't have that financial injection through to other industries. We rise and fall together - what happens when people made redundant are also consumers of your goods? It's a vital issue we haven't quite reconciled.
The human element
"We are creatures of little restraint who struggle with long-term planning. There's a concept in behavioural economics called 'hyperbolic discounting' - it's our inability to weigh up long-term opportunities against short-term value. Diets are a good example: You know what you want to be and look like in five years' time - and then there's a doughnut. The same is true of pensions.
"We know we should be putting more into our pension pots, but we don't because we want stuff now. And organisations have the same problem. Very few cast out 10 to 15 years to genuinely think about what is the best thing to do either in terms of societal or organisational good. Our inability to hold on to long-term goals is a challenge because the immediate challenges you face day to day won't be the ones coming down the road. So there will over the next few years be increasing automation and complexity of technological solutions, but it's very easy to sleepwalk towards them because it's never going to be your biggest issue in the morning."
Society, more broadly, is similarly afflicted with parliamentary terms of five years leading to potentially jarring swings in policy."
Changing attitudes towards data collection
D'Souza ties our more relaxed attitude to personal data use and how powerful it can be when aggregated by the firms collecting it.
"Predictive analytics is massive in the HR space, with a good example being the measurement of sentiment in our emails to tell whether you're becoming increasingly stressed and thus more likely to leave. Emailing while drunk or simply tired could lead to wild misdiagnoses or false positives, and a self-fulfilling prophecy when the individual involved eventually resigns out of frustration anyway.
"We need to be firmly in control of the technology, understanding how algorithms work and what they impact. We need to be confident that we are automating in the right way. It will be too easy over the coming years to simply bring in tech solutions because we're told it's the right thing to do. But you have a choice over how much you measure, how much you share that measurement, how much ownership individuals have and how much the organisation uses it."
Putting people first
"I'm delighted to see BIFM concentrating on the broader space beyond just the built environment, and at CIPD we're trying to do the same. We're thinking about the systems that make people more effective and productive.
"Dan Pink in his book Drive described the conditions required for productivity - autonomy, mastery and purpose. Autonomy is interesting because nobody wants to be reduced to just a number or a part in the machine; yet so much of what we're doing is reducing people to just that. We're asking - what category do you fit? How can I label you? How can I get an algorithm to predict your behaviour?
"Actually, one of the great things about us as individuals is that we should be unpredictable. That's one of the unique things about being people. By contrast, one of the most important things about machines is that they should be predictable and conform to rules. So if we're going to operate in an environment that's a blend of person and machine, what we need to nurture within that is our sense of creativity, uniqueness and unpredictability."
Yes, technological advances can influence productivity. But economist and former joint head of the United Kingdom's Government Economic Service, Vicky Pryce, opened ThinkFM conference by highlighting the broader economic issues influencing the UK's now famously low productivity growth.
Mentioning the squeeze in profits across various supply chains, and noting the events surrounding Carillion, Pryce talked of slow productivity growth being the result of a lack of investment. Figures show how the UK's annual average growth in output per hour had dramatically changed in the UK subsequent to the great market crash of 2008. Before this, the UK had vied with the US and Germany in productivity growth. It's since stuttered upwards at a far lower rate than the US, Germany, France and even Spain.
Brexit is difficult to avoid, with Pryce pointing out how construction, energy and business services were among 18 trade sectors identified by the UK government as having post-Brexit convergence with EU policy as a high priority. Soberingly, she also noted how no Brexit scenario showed the UK economy doing better in terms of gross value added (GVA) than it would within the EU.
Other indicators, such as the Purchasing Managers Index (PMI) showed first-quarter declines in growth and an actual fall in construction activity. All that said, unemployment remains at its lowest level since 1975, suggesting strong underlying conditions.
Sue Daley of Tech UK, a trade association for the UK tech sector, painted a vivid picture of the four main areas of activity influenced by AI and automation advances.
"There is a digital skills gap already with more jobs than skilled people to fill them. The pool of talented and skilled individuals will determine whether we can use these technologies."
"AI will creare new jobs, but many will also change and we need to think about what that means for our workforce and how we can retrain and reskill it so it is prepared and can act in this AI world. The opportunity is to free people to do more productive, enjoyable work."
"We need to build trust and confidence to bring everyone on the AI journey with us. This will involve ethical questions that go beyond the mere use of personal data. What does it mean to be human in an AI world? How do we ensure AI is working in the interest of humans and that humans flourishing is at the centre of tech development. GDPR is important and has been developed with machine learning, AI and autonomous systems in mind so the right legal and regulatory environment is in place but change moves so fast.
"We are encouraging people to think about how they can adopt, deploy and use AI technologies. The beauty of AI is that it will only be realised if we start using it. There are a lot of 'Terminator' style headlines in the news but we need to think about what this technology could do. If we are brave enough to embrace it, it will change our lives across the board, not just how we shop or work."