Customer centricity has been the main theme for any business model that Industrial sector is witnessing. Achieving customer centricity in a business to business (B2B) model warrants modernization of B2B operations, such as network consolidations to reduce costs and improve efficiencies. This increasingly demands connectivity to global manufacturing hubs located in matured and emerging markets. Flexible architecture, consolidation of ERP systems, and upgrade to “As a Service” business models are driving industry projects. Increasing demand of supply chain resilience and sustainability are demanding centralized command centers, control towers and predictive analysis.
At the end, the Internet of Things is causing a technological revolution and enterprises are seeking for experiential data to maintain profitably with evolving business and operating models stemming from the advanced IoT technologies. Disruptive innovation warrants precision. Operational challenges occur when machines fail to execute if timing is off by seconds. Synchronizing machine, people, and networks are key challenges and in turn opportunities, that will addressed in this white paper.
Themes of IOT and a Point of View (POV):
Internet of things is built of four major themes. They are:
• Sensors transactions
• Artificial intelligence and analytics
Each of these themes has its own unique value proposition, unique eco-system, and implementation strategy. In fact each of these themes individually is solving some unique business problems. The biggest challenge is to bring all of these four themes together and create one end to end solution for clients. Let us keep in mind that an organization (a client to a consulting firm) may still reap benefits of IOT with silo process transformations but the total cost of ownership is exponentially high in this approach. An organization should embrace a holistic view and consolidate hardware, software, security and implementation partners and products.
Manufacturing floors have witnessed sensors since 70’s. Why suddenly “sensors transactions” have got so much attention? Not that the infrastructure is new, it is about the maturity of the users which has made the difference. As industries have embraced automation, ask for autonomy has increased. To achieve autonomy, we need self-healing processes. To have self-healing processes, we need to re-define the boundaries of the partners in the eco-system. Sharing data is important for better decision making. Sensors industry have matured enough to handle huge data streams without becoming a maintenance hassle.
Data: Data assure facts. Facts drive decision making. Hence data and decision making defines transitive property of equality. Internet of Things cannot thrive without correct data. Effort to “trust the data” is the most critical journey in the implementation of Internet of Things and in turn Internet of Everything. As a consulting play, “data as a service” is becoming very prevalent. What large organizations like Nielson have done for consumer market, is on demand for all layers of transactions, be it business to business (B2B), business to consumer (B2C) and business to Enterprise (B2E).
Artificial Intelligence (AI):
Artificial Intelligence (AI) is about building systems or agents which have intelligence. These systems continuously learn and leverage a knowledge base. From a scientific and engineering perspective, Artificial Intelligence can be construed as a numerical recipe of any combination of algorithms and methodologies that help systems mimic human intelligence, identify hidden patterns, analyze data, learn business processes, aid decision making through text analytics and/or speech recognition. In a nutshell, AI is about creating SMART organizations that can predict and pre-empt certain business outcomes through machine learning and, presumably leverage historical data to its advantage. Plus, much more.
Artificial Intelligence & its increasing significance to businesses:
The pervasive deployment of IT in most business processes and workflows has essentially digitized enterprises, and continues to at an alarming rate. This has resulted in an explosion in the amount of data that is now readily available to businesses.
However, harnessing the full potential of this large volume of data that is now at our disposal is no trivial task. The real opportunity is to be able to extract meaningful insights and, equally important, take decisive action for competitive advantage – and, do it in an automated proactive way. AI driven analytics is, therefore, getting a lot of attention from business leaders and practitioners, alike. The following capabilities in AI are finding widespread use in multiple industry verticals:
Predict (or forecast) a certain outcome and, with what probability?
• Outcome could be any variable such as Demand, Sales, Failure Mode, supplier reliability, Customer Satisfaction, etc.
• Examples: How can I predict the ability of my business unit to meet the committed volumes of product shipments (essentially, YIELD prediction)?
Clustering/Segmentation of a population based on similar traits—Inferencing
• Understand Customer behavior/churn or how would I treat a new transaction (which cluster would it belong to, and therefore, what actions need to be planned, a priori)?
• Example: Which of my customers have a high propensity to switch to competition.
Extract hidden rules and pattern analysis (in conjunction with Clustering) to identify rules of Association
• Issue analysis, Return pattern Analysis, Predict likely outcomes based on hidden patterns/relationships (“if these …. then that”)
• Example: How may I mine my existing install base and effectively cross sell applicable services?
Text Analytics & “event” extraction using NLP on unstructured data
• Text analytics to infer events/activities that could potentially impact my client’s business
• Example: What is the overall sentiment of my consumer base through their conversations in various social sites and my own logs?
Impact Analysis (what-if analyses) including Optimization Models
• What would the outcome be, if certain state variables were to change? Or, how would I need to organize my enterprise so that certain objectives can get maximized (or minimized)?
• Example: How may I be able to isolate the root cause for a set of problems that I am facing in my organization? How may I correlate my defects to specific quality lapses in my engineering workflow?
• Example: How can I de-allocate my machine resources for predictive maintenance to ensure OEE
(Overall Equipment Efficiency) still remains within acceptable thresholds?
• What combination of product attributes would be ideal for the product to maximize its business effectiveness for the enterprise, taking into account multiple diverse dimensions?
Artificial Intelligence driven Analytics has immense potential in helping organizations predict outcomes, prescribe tactics, pre-empt events and institute a proactive approach in business operations.
To name a few examples, one could immediately see direct applicability of AI methodologies in Heavy machinery and shop floors in preventive maintenance of equipment and machinery. Or, in the area of customer experience by predicted the likelihood of customer churn or, for that matter, in proactively devising real time maneuvers to prevent dissatisfied customers from switching. In the chip manufacturing space, one could envision the deployment of AI to predict the yield in a chemical vapor deposition process based on the machine logs coming from multiple stations in process workflow. The list goes on.
Our focus in the current work is on Supply Chain and the various processes where AI has been successfully deployed, especially because there is a need to comprehend complex, interrelated processes, for which an intelligent knowledge base is critical for problem solving. As is obvious, the judicious selection of a specific AI algorithm, as a part of the overall numerical recipe, depends on a wide range of factors, namely nature of the “physics”, type of data, amount/sparsity of data, correlation (or lack of) between multiple variables, strength of certain associations, computational complexity/cost of specific algorithms, need for real time (or near real time) insights etc. etc.
Customers, suppliers, employees are demanding better experience in every interaction they have with business. This is adding to the pressure for change. It is a significant point of competitive advantage or disadvantage. Wholesale replacement of large parts of the IT landscape are unacceptably costly and risky. Reorganizing the IT landscape to separate a more stable set of core platforms from the more agile edge applications, offers means to deliver this change at significantly lower risk and cost. As we redefine business with plug and play mode of transactions, industry is witnessing a new business model for every aspect of the business. Be it merger, divestiture, operations, consumer retail, marketing, new business models with Internet of Things is a new normal state. To achieve total success, business accepts the new normal state where it finds simplicity, agility and scalability. Visualization defines the platform to these asks from business. Visualization not only depicts a running condition of the business, it also drives insights.
How Consulting can approach IOT:
Consulting, being the tip of the spear, brings thought leadership, change management, and business technology offerings in Industrial IOT (IIOT). Consulting, besides, core Engineering, presents a huge opportunity in this space. Offering such as “Servicification” of processes and technologies is brewing up like anything. Servicification can be delivered in three ways:
• Serving the client directly, completely private labelled, where customers of the client do not know the independent vendor to the client
• Serving the customers of the client directly along with the client with full non-disclosure and master service agreements
• Serving a client and in turn, bringing the client’s applications to another client of the consulting firm
All the models above can be commercialized or priced accordingly to the market standards. All the above models ultimately need to provide “As a Service” model where customers can pick and choose processes, technologies, integrations, platforms, etc. as and when they want to use them.
Industrial IOT (IIOT) and Manufacturing:
“Servicification” aka “As a Service” offering in Internet of Things maps people into a role based organization enabling better decision making with visualized data synthesized from captured data. “Servicification” offering in the industrial manufacturing industry vertical focuses on enabling new business models through “as a service” model for mergers, acquisitions, operations improvements, ERP in the cloud, etc. This offering is built on a culture of learning, creativity, and purpose. Its purpose is to renew the core and innovate the new.
“Servicification” targets to bring simplification and automation as well as user centric design and culture of innovation in the industrial manufacturing. Digital operations enablement is the key focus of this offering.
Process: This offering understands business processes, finds the cost of operations and recommends transformation in two steps:
• First, it will orchestrate the process areas by workflows
• Second, it will bring automation through sensors transactions wherever automation is feasible
The entire process optimization is achieved through weighted scoring measures and matrix.
Technology: This offering brings in productized solution. Any open sourced Information Platform delivers IIOT pre-packaged and brings “as a service” to our clients. Competitively priced, IIOT offering brings integration, data driven innovation and digital transformation leveraging intelligent machines and connected networks to collect data, analyze information and coordinate, execute actions in real-time. SaaS, DevOps supported by use of widely available—fast evolving Open Source platforms offer ways to deliver completely different kinds of solutions at completely different pace than in the past.
People: Role based approach embraces one key theme. Experience—customers, suppliers, employees are demanding better experience in every interaction they have with business. This is adding to the pressure for change. It is a significant point of competitive advantage or disadvantage. Wholesale replacement of large parts of the IT landscape are unacceptably costly and risky. Reorganizing the IT landscape to separate a more stable set of core platforms from the more agile edge applications, offers means to deliver this change at significantly lower risk and cost. IIOT offering delivers experience of visualization as well as experience of integration embracing Industry 4.0 framework at any level of organization.
In High-Tech industry, Micro-service offerings address this problem by integrating end-to-end business processes. These offerings provide business value, competitive advantage, and increased operational efficiency improving both enterprise top and bottom lines. Micro-service brings consumerization and productization by offering tailored solution to the client needs keeping standards in mind.
“Servicification” offering in the High-Tech manufacturing industry vertical focuses on enabling new business models through “as a service” concept for sales, services, product monetization, yield management, near-shoring, predictive maintenance, channel strategy for distributors, and centralized command centers. “Servicification” targets to bring simplification and automation as well as user centric design and culture of innovation in the High-tech manufacturing. Digital operations enablement is the key focus of this offering. This offering allows a client to customize its services.
Automotive Industry has witnessed a shift to “As a Service” model as well. Solutions like Telematics, Aftermarket, Infotainment, etc. are focusing towards productization and consumerization. Consumtpion management is one of the offerings that large automotive, and industrial organizations are delivering to their customers. Consumption management solution reads every equipment/cars via telematics solutions, parse the data and recommend insights to their customers for proper usage, insurance companies, and driver safety measures. Internet of Things services are becoming prevalent and easy to use for automotive industry. Automotive and in turn large rental organizations are experiencing the following asks from their customers.
Resources industry is another wing of overall manufacturing where Internet of Things is doing wonder. This industry is not only leveraging machines like drones to improve the crop growing mechanism but redefining the boundaries between the organizations and their growers.
Cloud as a Pre-requisite:
Prior to solving specific MFG industry problems via IOT, let us understand the importance of cloud in this play as well as what we need to prepare with to embrace “As a service” solution. Our clients realize the potential of having “everything” in cloud. This year, the market for cloud services is worth $43 Billion (source: Gartner). Yet clients and consulting firms are confronting a reality. Creating a new business model in cloud while keeping the existing models profitable, making the right calls to decommission a current model and transform into a new model is not in every firm’s “success” playbook.
To achieve success with new business model in cloud, there are a key prerequisite. As a firm we all need to ask one question to ourselves about what we are offering? We need to ask the same question to the clients we are consulting with and also to our partners. Are we clear on where our clients want to go in the next five years? Is our offering aligned to the client’s? Clarity is key, think big, start small, develop a prototype and get going. Once clarity is there, define a four step strategy to create the new business model for our clients. They are:
Step 1: Define a specific business model. For example, in Hi-Tech industry, whenever our clients are selling a “product”, they must sell a service along with it so that not only they go to market with an integrated offering but also to ensure that contract renewals happen pretty easily.
Step 2: Define a proper operating model. The operating model must consider four pillars. They are
• Delivery capability
Step 3: Define a proper model to co-exist non-cloud organizational blocks along with blocks which can move to cloud.
Step 4: Define a world class customer experience strategy. Our end goal is to offer a “service” bar to our clients. Allow them to pick and choose the services and enablers from the bar.
Solution Components for Supply Chain As A Service:
IOT Services warrants a significant consulting play to deliver various solution sets for an organization. Understanding the roadmap, aligning existing vs. new projects for an organization, persona driven business requirements mapping for future, IOT product benchmarking, etc. are very important steps prior to the actual implementation of IOT services. Solutions like Command Centers and Control Towers are allowing new supply chain servicifications. Following solution areas of Command Centers and Control Towers are direct needs of the MFG industry and its sub-vertical industries from supply chain management perspective where “As a service” (predictive analytics as a service to be precise) model is solving a ton of business problems.
1. Inventory Control and Management
Controlling the inventory and planning it at minimum cost is one of the major challenges in managing customer demand and expectations. The use of time histories from the past and the availability of real-time accurate information on expected customer demand forecasts, the size of inventory on-hand and the amount of order cycle time to fulfill an order can guide predictions and decision rules based mathematical models through AI.
Time histories of production schedules, BoMs and order patterns can be used by expert systems to estimate optimal levels of future orders, potential shortfalls and optimal timings of inventory replenishments.
2. Demand Planning and Forecasting
Accurate prediction of potential demand for products/services is extremely critical for every organization today so that the firm may plan out its deployment of manufacturing resources, hiring of manpower, scheduling of processes, new product development, market launch campaigns etc. A high degree of confidence in the ability to predict certain market events enable organizations to pre-emptively prepare themselves to capture competitive advantage or, for that matter, avoid launch pitfalls. Multiple forecasting techniques are used such as regression analyses, exponential smoothing, time history analyses, moving average, Box-Jenkins, ARIMA models etc
3. Order Picking Efficiency
One of the most time-consuming, labor intensive portion of warehouse management workflow is Order picking. The efficiency associated with Order picking is critical as it impacts the overall operating expense of warehousing. Artificial Intelligence algorithms are used in automatically optimizing the movement and deployment of manpower to maximize the number of orders picked. Also, in multiple instances at the plant floor, speeds of conveyor belts have been dynamically modified to meet sudden surges of order volumes.
4. Customer Relationship Management
Engaging customers on a continual and periodic basis is critical in nurturing long relationships to build trust and loyalty. Efficient hearing techniques have been devised to gauge consumer sentiment using text analytics. Proactive measures aimed at retaining customers, influencing their behavior and offering highly customized bundled offerings have used sophisticated machine learning/AI algorithms. In fact, effective cross selling and up-selling tactics around any combination of products and/or services leverage clustering and pattern matching algorithms within the overall AI recipe. Large firms have mined their Install base using Association mining techniques in order to maximize their revenue capture by better managing their entitlement frameworks.
5. Purchasing & Supply Management
Decision making in the context of a make-or-buy outcome relies heavily on insights through “what-if” scenarios and the business impact assessment. Mathematical models are built that are typically aimed at optimizing a set of Objective functions along with certain design variables and constraints to help guide the optimizer searches. Some typical use cases include business decisions around what products to produce versus procure, and in what volumes? What suppliers are more reliable for specific raw materials/products—based on their historical record? Or, what is the likelihood that specific products would pose a supply shortfall in the coming months? Etc. etc.
6. Logistics and Shipping
AI tools have been used extensively to better manage the logistics functions in order to reduce time for delivery or to reduce the shipping costs. Dynamically decisions are made to assign specific logistics partners for specific products based on their track record (reliability inferred through time histories). AI algorithms coupled with powerful Optimization algorithms have helped manage vehicle routing and scheduling issues, freight consolidation challenges and multi-modal connections. Genetic algorithms are quite popularly deployed.
7. Vendor Management
Almost every supply chain has a critical dependence on its eco-system of vendors and partners. The Supply eco-system could revolve around supply of raw materials, semi-finished goods and/or finished components. Additionally, there could be extensive eco-systems around Manufacturing as well as Logistics. Regardless of the type of eco-system, it is critical that the firm is cognizant of the reliability, performance and robustness of its partners/vendors.
AI techniques have been used to help negotiate vendor contract agreements based on the propensity of the partners to deliver committed targets in terms of volumes, timelines, costs, etc. Time history analyses helps define the rating of each vendor and also predict the likelihood of delays in shipment, for example when specific events get triggered. Also, correlation studies may be done to determine the dependence amongst certain conditions that could help identify the root cause for some undesirable outcomes.
8. Data Quality
At the heart of any workflow in any organization is the Quality of Data. With the pervasive deployment of IT across all major enterprise functions, tremendous amount of data has been gathered. No wonder, enterprise data is one of the key assets that an organization owns, based upon which useful queries and decision are made. However, the quality of data is an area of concern. Data Quality is measured along multiple dimensions—Accuracy, Completeness, Consistency, Compliance, departures from past nomrs etc. etc. Data quality issues may be attributed to multiple factors such as poor data capture sources, process discipline lapses, multiple system interchanges and simply the fact that certain data collections happen from the field staff that is continuously on the run.
Artificial Intelligence techniques have been used to address a large chunk of data quality issues. Supervised learning techniques using a training set of explicit quality signatures can be used in certain situations. Correlation studies, Time history analyses and nearest “neighbor” comparisons have been deployed to discern values for missing fields etc. AI has helped not only flag deviations from past norms or suspicious data fields, but has also automated the process of data correction through various sophisticated learning techniques.
The solution block diagram illustrates how supervised learning techniques may be used in helping fix data quality issues, as long as an explicit set of training data is available, for the most part. Additionally, time series analyses can be deployed to help identify patterns of departures from past norms. Based on the problem context, specific combinations of various methodologies could make sense.
Call to Action:
Consulting and IOT can go hand in hand. The notion of IOT being only an engineering play is no more valid. Here are few things where consulting can directly impact an organization’s appetite to embrace IOT and in turn Internet of Everything (IOE).
As we conclude, let us take each of the four theme of IOT and conclude with a view point.
We need to choose sensors very carefully. It can choke the production line from systems perspective. At the end, all data are going to cloud in this “As a Service” model. Choosing sensors and in turn defining compatibility of the sensors with the downstream machines will define the success of the overall implementation.
Creating proper data lakes and data governance programs are very important. Data as a service and in turn monetization of the data are becoming pure consulting play for consulting firms as well as Key Performance Indicators (KPIs) sitting on the reliance of data are becoming tools to sell services besides products for organizations.
The role for Artificial Intelligence is very significant in terms of the business impact that it potentially offers. Contextualization and conceptualization are critical parts of the overall AI/Analytics solution framework creation. It is critical to identify the specific Business challenges (thru customer interactions) while building a mathematical/numerical model to discretize the workflows/continuum. The conceptualized AI/Analytics solution also needs to tie in the process/engineering parameters with the business challenge. This is what makes AI succeed. Else, one ends up solving the wrong problem or the solution will never be relevant to the business problem, at hand. AI is a sophisticated numerical recipe of any combination of multiple methodologies and algorithms that needs to be deployed in the right context – for there is no silver bullet!
At the end, what we see, is what we believe. What we read drives our decision making. Visualization is very important. In “As a Service” model for consulting firms as well as for organizations delivering “As a Service” models to their customers (with the help of the consulting firms or not), visualization and depiction of data are crucial for overall success.
Anirban Bhattacharyya (Anir) has 17 years of Supply Chain (Operations) Strategy, IT Strategy and Product Strategy experience. He masters in advisory consulting focusing on benchmarking, providing thought leadership and implementing large physical and digital transformation solutions leveraging technologies like IBM, Amazon, Microsoft, Infor, Netsuite, Oracle, SAP and other specialized products.