Data-driven, AI-powered Supply chain: Part-3

Roadmap for Building a Data-driven Supply-chain / Supply chain 5.0.

9 min readDec 14, 2023

The immediate and pressing need for ‘digitizing’ your supply-chain

· According to a 2020 CGE report, companies which commit to digitizing their supply chains can expect to save up to 50 percent on supply-chain costs; besides, a 20% reduction in procurement costs, apart from an increase in revenue of over 10% — resulting from incremental competitive advantage. ( Source: Center for Global Enterprise (CGE) in partnership with CREATe.org)
· World-class supply chain organizations can save up to 45 % in procurement costs through digital transformation. (Source: Hackett Group)

One may conclude: ‘Digitizing’ the supply chain has become a survival necessity for companies to stay competitive. Apart from a substantial jump in the efficiency-effectiveness, the customer-experience, and upside to revenues, companies can expect a huge-huge cost-saving…

Components of Data-driven (Digital) Supply-chain

A good TOC practitioner, among other things, is expected to begin by describing the end goal… So, before we attempt to build a data-driven supply chain, let us first try to envision what it might look like.

Further, we may also note:

A data-driven supply chain is a supply chain management system that uses data analysis to optimize operations and improve decision-making processes. The features of a “data-driven supply chain” & the differences between the “traditional supply chain” and the “digital supply chain” are described in the figure(s) below:

Defining the “ideal” end-state: Featuring an AI-driven Digital Supply-chain of future

The supply chains of the future will have the following distinctive features: The emphasis would be on the following:

Figure-8: The Future of Supply-chain — Going 100% Digital

# Granular geo-tagged data from extended value-chain covering supplier’s suppliers; to customers’ customers

# Location-data (Location-stamp on the transaction),

# Robotics in factories and warehouses

# AI-driven algorithms for supply-chain planning and optimization

# AI-driven automated processes wherever possible

# Data-driven decision-making across the value chain.

# Self-driving trucks

# Internet Based Ai driven Bots (Listening Posts) for — Continuously scanning the market (internet) for better sources, better materials & better pricing & for quickly gathering & processing risk data (geo-political / supply-related)

1. Data Collection: This is the most important part of the ‘Digital Supply-chain’. The supply-chains of the future are expected to extensively use IOT, Radio-Tags, GPS, etc. for Automated Data Acquisition (event-driven, or period-driven in rare-cases).
# Enterprise data enhanced by seamlessly integrated geo-tagged micro-market data (perhaps made available on subscription mode by a third party)
# All transactional data without exception will be geo-tagged & instantly distributed to cloud-based Apps & Data lake-houses.
# Every Truck would have a GPS its location will be tracked 24x7. Every “Distribution Package” will carry a Radio-tag and its movement will be tracked through supply-chain.
# Every purchase, every order fulfillment, every job progress, every project will be tracked with time & location-stamped data.
# Every product & Every Job would carry a Radio-tag, and its movement through the workflow-stages would be automatically tracked.
# Every Store & Warehouse would have completely automated, AI-driven — Automated Storage and Retrieval Systems (ASRS), that automatically load and unload the unitized-palletized loads, while keeping track of the locations (Bin Number) for every Part-No & every SKU.
# All equipment & devices without exception would be IOT/GPS enabled and would automatically generate and distribute data into multiple analytics engines for control & continuous improvement.
# All workers may carry id-tags with a GPS, or their mobiles will be tracked through a custom-corporate app.
# All intra-office communication on mobiles through short-messaging apps & e-mails will be time-stamped and location-stamped.
# Real-time Batch Tracking & Product Traceability data through Radio-tags & Unique-ids for each SKU
# All communications with customers and vendors will be time and location-stamped.
# The internal Enterprise-data will be complimented/supplemented with external-data, from sources such as Census data, Market data, etc., — all sorted, normalized, and geo-tagged, before uploading into Data Lake-house.
2. Transportation:
# Every Truck would be Self-driven, GPS enabled, and will have the facilities for sending emergency distress signals.
# Digital Twin (SCADA equivalent system to control Transportation network) to carry drill-down information of every package, every SKU carried by every Truck, down to the level of ‘serial number’ of every SKU.
# AI-controlled Self-driving Drones for delivering products — especially for the last-mile delivery.
3. Warehousing & Supply-chain Operations: Completely Robotized — Automated AI-driven –
# Robotized picking and packing, AI-driven automated storage and retrieval system (ASRS)
# Automated Instant-Inventory taking at the press of a button using radiotags on every box, every pallet.
4. Customer / Vendor Collaboration
# Generative AI-based Chatbots to communicate
# Generative AI-based IVRS & Robotic calls for voice communication
# Generative AI-based Social Media Campaigns
5. Risk Management :
# Risk data to be collected using AI-driven Bot Listeners for picking up “Risk Signals”, “Mavericks”, or any “unusual activity”, or “unusual communications” — both from internal-enterprise-data & from the other external-data-sources including Public & Private databases, Social Media, and the Internet.
# AI-driven Risk Reporting, Early-Warning-Signals
# AI-driven automated ‘course-correction’ and ‘communications’ with vendors, customers, distributors, and employees.
# The Lifed-items (SKUs with an expiry-date) to be picked and packed based on shorted expiry Batches first.
6. Product Traceability, Batch Tracking & Adverse-event Reporting
# Real-time identification (Batch No & Serial No.) & Location tracking of product through Radio-tags & Unique-ids for each SKU — which maps to product-traceability-data on the cloud
# Traceability data includes: the origin of materials and parts, their processing history, and their distribution, as stated in ISO 9000:2005 (2005) & Additionally, data connected to the past, use, or location of a product.
# Easy & Instantaneous Batch-Recall in case of any AER- Adverse Event Reported (as in the case of Food or Pharmaceuticals, or even consumer-durables like automobiles ) with reference to a particular batch-number, would automatically trigger an alert along with a “Do not Sell / Do not Use Warning” — not just for the particular batch-number, but also for all the batches which used materials & parts from the same input-batches from same suppliers. The alert would automatically replicate itself through the entire supply-chain (distribution network) no matter where the particular ‘risky’ batches are located.
7. Digital Twins:
# To be Mobile based: To be operated from “Anywhere & Anytime” for Planning-monitoring-correction & control
# Digital Twins to map 100% of the extended-supply-chain, all stages, including the controllable Vendor’s Supply-chain & Customer’s Supply-chain.
# Control Tower & AI to Co-Pilot decision making: AI-driven control tower to help managers make decisions & manage the overall performance of the supply-chain. Managers handle extreme exceptions alone.
8. Global Sourcing & Continuous improvement
# Continuous & Automated Scanning of the Internet for better Suppliers, Materials & better prices — using Bots
# Use of Generative AI for Price-Negotiations with customers.
# AI-driven BOTs (Listening-posts) on the Internet, and Social-Media for picking-up signals on Threats (such as Adverse Events, Geo-political risks), or Opportunities (better materials, better prices, better sources

Building a Data-driven Supply-chain

As someone familiar with my Data Science Central articles, or my recent book would know: I have extensively written on data-driven decision-making, data-to-decision lifecycle, and creating a roadmap and a business case for building a data-driven organization earlier. (Big Data for Big Decisions: Building a Data-driven Organization, P. Taylor & Francis, UK, 2022).

As I have argued: fancy software platforms, AI, and analytics apart, the key to building a Data-driven organization is to focus on the key decisions — the 10% of the organizational decisions that account for 90% of business outcomes.

So for building a data-driven supply chain, the following key questions must be asked and answered:

1. What are the key “supply-chain decisions” that account for over 80–90% of Supply-chain efficiency?

2. What is the data that supports the identified key decisions?

Note: The bigger and more primal issue most organizations face is: that they simply do not have the data that supports (the) key decisions within the organization…not with the requisite granularity or the quality? Needless to say, building a data-driven supply chain in the absence of the data is a futile and vain exercise, which nevertheless many CDOs are attempting-to.

To quickly summarize, I advocate the following steps (Refer to Table below) for building a data-driven supply-chain

Roadmap for Building A Data-Driven Organization

Building A Digital Supply-Chain

1. Start from decisions… List all Organizational decisions, along with metrics like “Value-at-stake”, and “frequency”

How do we list the decisions?
# Create a Master-list of Process-Constraints & List decisions associated with each process constraint.
# Each process constraint at each workflow-stage represents a set-of decisions (There can be many process constraints, but only a few significant ones)
# The same/similar process constraints may be recurring at multiple workflow-stages… Note the Frequency of recurrence.
# Working out “value-at-stake” for each decision: Can we estimate the ‘loss of throughput’ due to bottlenecks, at each stage?

2. Identify the 10% of the decisions that account for 90% of business outcomes

# Can we list the bottlenecks — from the biggest to the smallest?
# And the 10% of the bottlenecks that account for 90% of throughput loss?
# Create a List of Process-constraints associated with the 10% of ‘critical’ bottlenecks.
# Note: All workflow stages are important. The absence of a current bottleneck does not diminish the possibility of a future bottleneck.

3. Create a Roadmap & Business case

# Create a roadmap for digitizing the supply-chain by prioritizing the Process-Constraints associated with the 10% of bottlenecks responsible for 80–90% of throughput loss.
# Note: Digitizing the key process constraints at all workflow stages is important… The criticality of a bottleneck (& associated process constraints) dictates its priority.

4. Data behind the Decisions

Create & Map the Information Supply-chain
& the following as needed:
# Services Supply-chain
# Cashflow Supply-chain
# Map the flow of information (Data + Insights)’ required at each workflow-stage
# Map the flow of information (Data + insights) required for ‘easing out’ the process constraints and the possible bottlenecks.
# Map the Information supply-chain in its entirety — (from data creation to distribution, consumption, archiving, retrieval, and repurposing)
Cross-map the Information Supply-chain with Physical-goods Supply-chain — (clearly indicating the cause & effect relationship between the ‘information availability’ and its ‘impact on the process-constraint’.)

5. Data that you need vs. data you have

Check:
# If you have all the data that you need to support the Supply-chain decisions
# For the Delta-Data: Create & implement a Data sourcing strategy

6. Building Analytics to improve the quality of decisions

# Build Analytics & AI solutions to improve the quality of Supply-chain decisions

7. Build a Digital Twin:

A remote replica that provides the status & the control mechanism for the entire supply-chain (similar to SCADA systems). Typically you can change the process parameters using ‘digital twin’.

READ EARLIER:

References

1. Parra, X., Tort-Martorell, X., Alvarez-Gomez, F. et al. Chronological Evolution of the Information-Driven Decision-Making Process (1950–2020). J Knowl Econ (2022). https://doi.org/10.1007/s13132-022-00917-y

2. Leigh Buchanan & Andrew O’Connell. A Brief History of Decision Making, HBR Magazine, Jan 2006. https://hbr.org/2006/01/a-brief-history-of-decision-making

3. Power, D.J. A Brief History of Decision Support Systems. DSSResources.COM, World Wide Web, version 4.0, March 10, 2007. http://DSSResources.COM/history/dsshistory.html

4. H. Steyn. Project management applications of the theory of constraints beyond critical chain scheduling. International Journal of Project Management 20(1):75–80, January 2002. DOI:10.1016/S0263–7863(00)00054–5

5. Dr. Eliyahu Goldratt. Standing on the shoulders of Giants Production concepts versus production applications The Hitachi Tool Engineering Example. 2006. https://cdn.ymaws.com/www.tocico.org/resource/resmgr/standing_on_the_shoulder_of_giants/standing_on_the_shoulders_of.pdf