13 Aug A Unifying Framework for Digital in Business
Is there a unifying model to help decision makers sift their way through the digital echo chamber? Here’s my version.
For the past 18 months, as I’ve been working at the intersection of Digital and Oil and Gas, I have been studying the leading digital technology categories, the various stages of the oil and gas value chain (upstream, midstream, downstream), and the visible interplay between these two domains. Throughout my study, certain patterns keep repeating themselves—sensors generating so much data that they outstrip human-level tools to analyse, robots working optimally with high quality operational and commercial data, and cloud computing environments appearing virtually everywhere in tandem with digital innovation.
A Unifying Model
In hindsight, I had been hunting for a unifying model or theory that tied these digital innovations together, predicted how these technologies behaved, and forecasted how they likely evolve. Woven throughout the model is a story about data—generating, analysing, consuming, managing and presenting data. Data is not generally viewed in Oil and Gas as a corporate resource on par with other assets. Data management receives little attention, and roles in data science are only just beginning to surface. Accountability for data in Oil and Gas is diffused and fragmented. But winning at digital means winning at data.
The Internet of Things Generates The Data
The first element in the model is the sensors and the rise of the internet of things. Sensors generate enormous quantities of data, with greater diversity in form and content, at ever declining costs. They are appearing on virtually everything—pumps, valves, vehicles, vessels, and people. Oil and Gas, as an asset intense industry, drives demand for internet-connected things.
The trajectory is pretty clear—smaller, cheaper, more capacity, lower power, encrypted, connected. More smarts embedded in the sensors means these devices will be able to do more and take on more local work. Instead of ten unique sensors on a thing, there will only be one super-capable sensor.
AI Interprets the Data
Only the modern tools of machine learning and artificial intelligence are able to process the immense volumes of data that the sensors generate. The breakthrough technologies of the 1990’s—spreadsheets, personal computers— are not up to the task of storing, manipulating and analyzing the rapidly rising tide of data. Spreadsheets are not going away, but they are no match for the kinds of processing needs we now face. AI drives demand for data science professionals, and is a key reason why universities and technical schools are rapidly revamping their training programs to incorporate more emphasis on the data professions.
The amount of money and investment pouring into AI, coupled with the phenomenon of fleet learning—individual AI engines that share what each other learn the instant they learn it—point to constantly falling cost and improving capability. Eventually, job design starts with AI and incorporates the human attributes, rather than the other way round.
Robots Apply the Data
Autonomous technologies (or robots) consume the AI-interpreted data to carry out real work. Heavy haulers in the oil sands mines are good examples—their onboard cameras and sensors feed data into AI machines to interpret the real world and its hazards, and the hauler starts, stops, turns and accelerates. For the moment, a human controller is manipulating these machines, but in time even that task becomes unnecessary.
Robots are not confined to the field. Increasingly robots are office-bound, and others exist as services available in the cloud accessed through a web browser. In the future, many industrial machines, like air conditioners, water pumps, sprinklers and compressors will have on-board AI capabilities through which they will make increasingly smart decisions. These decisions include, given a set of changing conditions, when to run, at what pace and using what resources at what cost. Today, only humans make these decisions, or don’t make them at all.
Cloud Stores the Data
Cloud computing is the logical place to store all the data and house the analytics. Companies can use their own proprietary storage and computer systems but cloud is much faster to deploy, is more secure, and can ramp up and down more easily. The volume of data in play suggests that a cloud strategy is pretty much a one way street—there’s no reversing course.
For technology providers, cloud is very compelling. Instead of developing for multiple database environments, they select one, on which they become very capable. Upgrades and patch releases are executed with agility—no more years’ long upgrade cycles. Amazon is said to release millions of software updates a year, something not possible if Amazon’s software is installed on customer premises. Software distribution is via download from stores, not shipped on CDs. Users subscribe to services, rather than obtain licenses to use, a model that is much easier for businesses to control.
All those sensors need the occasional (or frequent) patches to deal with viruses and attacks, too. The only practical and cost effective way to maintain software reliability on these sensors is through cloud-enabled subscription and distribution. Cloud contracts are very easy to acquire, and are often embedded inside service agreements. Companies find themselves with hundreds of cloud enabled solutions over time, driving interest in cloud interoperability, data standards and integration between environments.
The end of the traditional software business model, which has many implications for how companies govern their use of technology, is rapidly approaching.
Blockchain Provides Trust
The rapid growth of sensor technology, robots and analytics drives the need for a new trust mechanism that is not human centric—blockchain. Blockchain technology monitors the state of the sensors, data movements, the AI engines and the robots to assure they are reliable and not compromised. By assuring trust, blockchain confers agency on the robots, allowing them to operate without human supervision or human intervention. This has spill-over effects on traditional people-based business processes such as loyalty—robots with agency may not be influenced by loyalty schemes.
Blockchain opens up entirely new business models, such as asset sharing. Instead of a business needing to own an asset, it can subscribe to the asset, and pay only for the cycles consumed, which are recorded with trust on blockchain. Balance sheets are transformed when long life, low utilization assets can be available when needed.
ERP Enables the Commercial Environment
Enterprise Resource Planning systems are themselves becoming digital and are a key part of the future where they support the commercial processes of buying, selling, tracking and measuring. In time blockchain technology displaces some commercial functions too, but for the time being, ERP is the commercial backbone.
ERP systems now embed AI engines and blockchain support within, which makes them potentially much more useful across a broader sweep of the organization. Integrating operational assets with commercial ERP data is a defining feature of digital—a compressor uses its own on-board AI capabilities to decide which power source to use based on incoming weather, and the market price of power, contract for that power, and pay for it as a single integrated transaction.
Agile Methods Are How Digital Gets Done
Digital innovation requires faster ways of getting things built and deployed, and Agile is the language and method of streamlined work practices. Hand in hand with Agile are faster ways to introduce change to operating environments (or DevOps) and better ways of interacting with technology (or User eXperience).
The logic of separate technology organizations (one for commercial technology that delivers business change in one fashion, and a different organization for operational technology and managed in a different fashion) falls away. IT and OT need to work together to deliver digital solutions and need a common way of working. Agile is the basis for that collaboration. Industry leaders merge their IT and OT organizations into one unit with one executive accountability.
People Manage Change
The future of work for people is designing and building digital environments that integrate these digital innovations. Multi-disciplinary teams that assemble new methods of working, deep data skills, process modelers, system integrators and business model designers, work together to construct new digital solutions using Agile methods. Helping people cope with change relies on our human only skills of creativity, empathy, story-telling, teamwork and problem-solving.
Business needs both new kinds of talent and purposeful re-skilling of existing talent for a digital world as the methods of working—stage gate, asset-centric—give way to Agile and data-centric. For example, robots require their own handlers, AI needs mathematicians, and the mountain of data calls out for data scientists. Dozens of unfilled jobs for individuals with these skills are listed on the many job sites, not just for Oil and Gas, but in many other industries that are experiencing the same digital wave.
So far, I’ve found this model to be very helpful in explaining the interplay between the various features of a digital world and how they work together. I hope you do too.